*2.4. Data Analysis*

The effect on springtails life-form groups and QBS-c index was analyzed with the mixed model in SAS University Edition (proc Mixed). In the analysis, the effect of crop and treatment, as well their interaction, were included. The year and term of the study were the random factors. The abundance of springtails per sample was relatively low (with the number of individuals of 10.5 per sample). Therefore, the abundance of springtails was calculated for 1 m2, knowing that the area of one sample was 0.000785 m<sup>2</sup> (5 cm diameter).

The springtails abundance as well morphometric trails in relation to experimental treatments was analyzed using canonical correspondence analysis (CCA) in Canoco, Version 4.5 (Ithaca, New York, USA). Significance of the first canonical axis and all axes together was calculated with Monte Carlo Test.

#### **3. Results**

The abundance of Collembola per m<sup>2</sup> and the QBS-c index differed significantly between all tested factors (Table 4). Generally, the most abundant group was epedaphic, then hemiedaphic, and the least, euedaphic Collembola (Figure 1). Within the epedaphic and hemiedaphic groups, both the plant (p < 0.0001) and biochar (p = 0.0009, 0.0058) significantly differed with respect to springtails abundance. Springtails were significantly more abundant in oilseed rape in comparison to maize crop. At the same time more Collembola were found in biochar, but only in oilseed rape. The abundance of euedaphic Collembola differed between biochar and control plots (p = 0.003). In both crops, significantly more individuals were found in biochar treated soil.

**Table 4.** Results of repeated ANOVA (GML, p ≤ 0.05) considering effects on treatment, plant and year, and its interactive effects on Collembola life-form groups and QBS-c index.


\* F = ratio of two different measure of variance for the data.

**Figure 1.** Effect of crop and biochar application on different Collembola life-form groups. Note: (**a**), (**b**), (**c**); indicate significant differences (p ≤ 0.05) between biochar and control; A, B indicate significant differences (p ≤ 0.05) between plants; the results of repeated ANOVA used for data analysis are given in the Table 4.

Considering the QBS-c index, it was significantly higher in oilseed rape crop in comparison to maize (Figure 2) (p = 0.0292). In both plants, the index was significantly higher in treatments, where biochar was applied (p = 0.01132). As shown by the life-form groups, significantly higher QBS-c index was found in oilseed rape only.

**Figure 2.** Effect of crop and biochar application on the QBS-c index. Note: (a), (b), (c); indicate significant differences (p ≤ 0.05) between biochar and control; A, B indicate significant differences (p ≤ 0.05) between plants; the results of repeated ANOVA used for data analysis are given in the Table 4.

The morphometric traits used for the calculation of QBS-C index were correlated with experimental treatments (Figure 3). The significance of the first canonical axis (CCA1), as well all axes together was p = 0.002 (Table 5). Biochar (in oilseed rape) was positively correlated with reduced legs, antennae and furcula, as well absence of specific structures on cuticle. In maize, where biochar was applied, springtails were characterized by a reduced number of ocelli. Size and pigmentation were positively correlated with oilseed rape and maize, both without biochar.

**Figure 3.** Canonical correspondence analysis (CCA) biplot on Collembola morphomentric traits in relation to experimental treatments. Notes: The morphometric traits are described in more detail in Table 1.



The eigenvalues of the first two CCA axes were 0.0102 and 0.041, respectively (Table 6). Both the first canonical axis (CCA1), as well all axes were significant (p = 0.002, Monte Carlo test). As shown on the CCA biplot (Figure 4), the Collembola community was affected more by crop than by biochar. There was only minor effect of biochar in oilseed. In maize the group of species related to the control site (e.g., *Desoria tigrina* and *Pseusinella sexoculata*) differed from the species which preferred biochar (e.g., *Stenaphorura* spp., *Isotoma antenalis*). Considering life-form groups, most of the hemiedaphic species were found in the oilseed crops to oil seed (with similar effect for biochar and control). The epedaphic species were frequently found in all of the treatments, while euedaphic mostly in maize with biochar.

**Table 6.** Results of Canonical correspondence analysis (CCA) of Collembola species in relation to experimental treatment.


**Figure 4.** Canonical correspondence analysis (CCA) biplot on Collembola species in relations to the crop and biochar application. \* Designation of the life-form groups: epedaphic, hemiedaphic, euedaphic. Note: the detailed description of the species is included in Table 1.

#### **4. Discussion**

The low temperature pine/wood chip biochar used in the experiment is characterized by the high carbon content but low surface area, and ability for nutrient storage, therefore the effect of its application to soil had little effect on soil properties. Tested pined wood biochar, was free from PAH and the concentration of all tested toxic compounds was very low or even under the level of detection. The release of toxic compounds like PAH's and heavy metals after biochar application seems to be crucial for survival and reproduction of soil fauna [47,48]. Some of the soil properties were improved, like total organic carbon content (only in maize trials) or CEC, as reported by other authors [49–53]. The liming effect, which is mostly expected [54,55], when biochar is applied to soil was not determined in our experimental trials.

Our estimations showed a higher QBS-c index value and higher individual number of particular life-form groups in biochar amended soils, confirming the hypothesis of soil biological properties improvement. We state that the positive response of Collembola to biochar addition was the result of improved soil properties. Some authors have found that soil mesofauna abundances are closely related to soil conditions, especially soil pH and organic matter content [56,57]. To compare, the significant increase of fungivorous nematodes was found after biochar application in the rates from 12 to 48 t/ha [26]. In contrast, no response of soil faunal feeding activity to biochar addition was found in the rates range from 0 to 30 t/ha [58]. Also, Castracani et al. [59] did not find any interaction between biochar and the abundance of epigeic macroarthropods in the rate of 14 t/ha.

The response of euedaphic springtails was predicted to be most distinct after biochar amendment. This would result from low dispersal ability and living in deeper soil layers [60]. Therefore, euedaphic springtails would be more sensitive to changes in the soil environment [61]. In our experiment the abundance of euedaphic springtails was relatively low compared to hemiedaphic and epedaphic groups. However, in both analyzed crops, its number increased after biochar application. Considering the two other groups living on upper soil layers, the effect was significant only for oilseed.

The main concept of the QBS-c index is that soil quality is positively correlated with the number of Collembola species that are well adapted to soil habitats [17]. Otherwise, numerous occurrences of euedaphic forms of springtails in a given habitat can indicate the better biological quality of the soil [43]. Based on the results, we can agree with the hypothesis that the QBS-c index will have higher values in crops where biochar was applied. For instance, in the study of Twardowski et al. [61], the QBS-c showed higher soil quality in the potato crop rotation in comparison to monoculture. Jacomini et al. [62] found decreased QBS-c index values in degraded soils. To confirm our last hypothesis, two springtails life-form groups (epigeic and hemiedaphic) differed significantly between crops. More springtails were found in oilseed rape in comparison to maize crop. Considering that maize and oilseed rape are plants which differ in their development, this should also affect organisms living on the soil surface. Similarly, Op Akkerhuis [63] found differences in mesofauna abundance between crops, i.e., specifically the mesofauna groups were much more abundant in cereals than in root and tuber crops. We state also that cover plant might modify the effect of biochar on soil fauna.
