The easiest and most repeatable substrate for field-scale purposes was identified as being a hay for small animals (“Vita Verde Small Animal Alpine Hay”, by Vitakraft pet care GmbH & Co. KG, Bremen, DE). The hay was ground in a 3 mm grid forage mill (Retsch GmbH, Haan, DE). About 5 g of hay was packed into half empty 10x10cm square polypropylene nets (1.5 mm mesh), which were then resealed using 4 staples, and a plastic label was added for identification and for easiness of finding purposes. The probes were inserted vertically near the roots and remained underground for about 60 days. They were then dried at mild temperature, gently cleaned and preserved at room temperature.

The chemical composition of the hay and litter-bag residues was predicted, using a Perkin Elmer IdentiCheck ^TM instrument (714-3333 nm), and the used equations were established on twelve species of crops, analysed at four stages, as reported by Tassone et al (2014) 29.

The extracted litter-bags were opened, and the surfaces of both sides were examined using a smart new miniaturized NIR web-based wireless spectrophotometer (SCIO v. 1.2, Consumer Physics, Tel Aviv) with a 740-1070 nm range. Three spectra were acquired on the two sides of the litter-bags. Chemometrics of the 331-point spectra was performed using a categorical discrimination procedure, integrated within the SCIO Lab proprietary software named AKA (Also Known As), and the confusion matrix, after normalization and 1^th derivation of the spectra. The reclassification capacities in the Symbiotic (S) and Conventional (C) classes within each calibration experiment, where S litterbags were compared with C litterbags, or in validation experiments, with only S litterbags, were considered as the reference of the performances.

The calibration of the SCIO spectra from experiments 1-8 is reported in Table 2. The average AKA reclassifications were 71% for S (P<0.0001) and 62% for C (P<0.0001) , with variation coefficients of about 25% between experiments. The results were confirmed from the validation sets in experiments 9 and 10 (Table 2), where overall classification ratings of 78±4% for year A and 71±5% for year B were obtained. Among the eight models, numbers 2, 3 and 5 were the best ranking ones for the A and also for the B years : in fact, their average classification ability was 89.9±3.1% and 90.1±3.6%, respectively. Compounding information from the best three chained models (Table 3) raised the probability of not obtaining one false negatives in a correct classification for true S membership to 98.4±0.12%.

The overall average validation grade was 2.73±0.25 (data not shown in the table), a sure sign of effective modifications in the litter-bag composition after the BMS treatments.

As far as the evolution trend of the litter-bags (Table 4), compared to the original hay, is concerned, the overall result of the degradative processes increased the value of the multivariate crop maturity index towards a more mature type of forage, by 81% in C and 64% in S. The percentage of recalcitrant components increased in the litter-bags: hemicellulose (88% C and 98% S), ether extract (55 and 60%), ash (45 and 47%), indigestible NDF (26 and 9%) and lignine (ADL 17 and 19%). On the other hand, the labile components underwent an average relative decrease: crude fibre (-22 and -36%) and acid detergent fibre (ADF -19 and -22%). As for the MBS inoculation (Table 4), five significant variables distinguished the S litter-bags from the C ones: the nitrogen-free (NFE -3%, P=0.007) and the wall components decreased (NDF -10%, P=0.0006; indigestible NDF -13%, P=0.08 ; crude fibre -17%, P=0.0541), but the crude protein increased (+13%, P=0.0025), and was thus apparently more protected from the added BM. Moreover, the lipids increased (+6%, P= 0.0247).

According to Tassone et al. (2014) 29 concerning the algebraic formula of the crop maturity index for growing plants, the regression sign of the percentage on the days from sowing was positive for NDF, ADF and indigestible NDF, while it was negative for ash, crude protein, NDF digestibility and digestible-NDF. After haymaking, in the underground environment the grasses composing the litter-bags started an ontogeny involution, as a result of biotic and abiotic factors, but also because of BMS action. The observed rise in protein may be a sign of increased MBS growth 25, 31 , and the net result was that the fibrolytic communities elicited the attacks of the carbohydrates. In terms of crop maturity index, the BMS increased the evolution of the litter-bags towards a more mature type of residue, as can be observed in Figure 1, where the lines of the S and C trends cross. Our results suggest that, as expected, BMS promotes the mechanisms that are favourable for an early maturation of the residual organic matter in the root horizon, and multi-annual observations are necessary 32. BMS management is based on the inoculation of aerobic microbes, but, because of a luxuriating rhizosphere, and in spite of respiration-fermentation processes, the net long term result could improve the carbon footprint of the whole plant-soil system, and thus raise its sustainability. Several soil management practices, inspired by a conservative agriculture design for the improvement of the accumulation of soil organic matter, are largely supported by EU agricultural policies, while the pro-MBS route has been totally neglected, in spite of the phosphate crisis that is expected from 2030 onwards 33.

The results of the present work are in agreement with Leolini’s 34 results (ibid Table 12) obtained from a litter-bag quality study conducted on six natural sites in Spain, in which four litter types were considered. In fact, as a result of some mild differentiated activities, a significant linear increase was observed in the Spanish study for the ash and ether extract percentages, but the digestible NDF also increased, while no labile constituent was reduced. In the agrarian and well cultivated soils of the present work, the litter quality appeared to be modified to a great extent, and the MBS attacked the labile components, which arithmetically enhanced the percentages of the counterparts. In the MOLTE long-term organic experiment 26(ibid Fig. 72), only a mild footprint of the Organic vs. Conventional soil was found by means of a multivariate analysis of the predicted constituents (R²0.18±0.04) or, rather, in the direct NIR spectra discrimination (R2 0.32±0.06).

Litter-bag decomposability, intended as mass decay, can be related to NIR spectra 35, and these correlations were also present in the data from Florence University: in 34 R²0.46 for the residual mass; in 28 0.74 for the residual mass and 0.90 for the k decay term (recalculated). In the present work, the quantitative aspect of litter-bags was omitted, because of operational difficulties at a large scale, but also after results from the MOLTE experiment which elicited a more meaningful structural and functional relationship from the variability of the litter-bag quality than from considering the total lost mass or the exponential decay. MBS activities in litter-bag matrices mainly depend on the unexplored vast communities in the foreign soil, and to a lesser extent on the microbiome of the hay. The outcomes of litter-bag modifications are also modulated by abiotic factors, such as the pabulum conditions, i.e. thermic, water and mainly the redox-oxygen availability. According to 36, an incubation of litter-bags for two months could allow the net N mineralization to be estimated, and in the present BMS framework, an N preservation appeared. The effect of the MBS treatment on the symbiotic fingerprint appeared to be quite consistent and repeatable for the five significantly varied constituents.

As far as the discrimination problem is concerned, can these five signs be considered a valid support to obtain an univocal response that could help to testify the use of MBS as biofertilizers? The chemical composition of litter-bags needs a chemometric deconvolution of a broad NIR-IR spectra (714-3333 nm), which could be obtained from high-quality scientific instruments. These devices represent a valid tool to help understand some mechanisms, but are less portable for a large-scale dimension. Thus, thanks to the overtones and combinations of the organic molecules in the electromagnetic spectrum, originating in the IR region, a surprisingly small but rich NIR spectra 740-1070 nm (Figure 2) can be capitalized on by means of vibrational spectroscopy. For field sampling and analyses operations on a smart-farm basis, the S footprint should be directly searched for in the electromagnetic spectrum. Considering the immensity of the biotas in different farms and crops, a rational choice among local models could protect against gross biases. The between-farm validation adopted in the present work is similar to a local vs. global chemometric procedure, utilized to manage large NIR datasets of soils in a better way 23. The false negative litter-bags, with an S grade of 1, were mostly concentrated on two farms. This may have been the result of a real inefficacy of the BMS for those particular management conditions.

The outlooks on the use of NIRS regard both plant tissues and canopies 37, 38 as well as soil quality for precision agriculture purposes 39, all of which require new approaches to acquire soil data on landscapes 40, 23. Direct NIR scanning of the soil horizons has also been proposed as a valid and practical tool to monitor the ontogeny and heterogeneity of detritus in soil, which is useful for the assessment of the carbon and nitrogen budget of the soil 41, but even for the macro components of soil biota: in fact, Zormoza et al. 42found very high r-squares for AMF (0.91), Fungi (0.80), Protozoa (0.73), Actinomycetes (0.92) and Bacteria Gram+ (0.91), and also for enzymatic activities, while direct NIR scanning was not so reliable for exchangeable P (0.46).

The authors wish to thank: S. Capaldo and M. Seminara (Consorzio La Granda Quality, Fossano, CN); R. Bodrero (Commerciale Agricola, Villafalletto, CN); G.L. Malvicini and G.L. Turello (illycaffè S.p.A., Green Coffee Procurement Dept., Trieste); A. Bevilacqua (AGRION, Manta, CN); G. and M. Bergese (La Corte, Monasterolo di Savigliano, CN); R. Del Negro (Aix en Provence); S. Ravaglia (SIS Foraggere, BO); F. and R. Polo (ATS. Bio contrasto al CoDIRO, Ugento, LE) for their kind support in the setting up of the experiments and in the litter-bag management; thanks are also due to the Fondazione CRT, Torino for the financial support to the scientific activities of the Accademia di Agricoltura di Torino and special thanks to L. Leolini and M. Santoni for their pioneering work.