3.1. Experiment 1—Chemical Composition
The chemical composition and phenol content of the individual by-products prior to ensiling, used in experiment 1 and experiment 2, are listed in
Table 2. In particular, OMWW and GP are characterized by phenol content (37.5 and 11.40 mg GAE g
−1 DM), and CW by protein content (11.30% DM).
The variations in the chemical characteristics of SIL-A and SIL-B are presented in
Table 3. Before ensiling (D0), as expected, SIL-B had greater (
p < 0.05) dry matter in comparison to SIL-A silage, which showed higher (
p < 0.05) crude protein, ether extract and ADL contents. This was determined by the different amounts of the by-products used in the silage formulations, mainly straw at 60% or 40% and grape pomace at 20% or 0%, in SIL-A and SIL-B, respectively. This affected the composition of the two silages in terms of their individual chemical characteristics.
Similarly, at D30, the SIL-B was characterized by a higher (
p < 0.05) dry matter content, while SIL-A had a higher (
p < 0.05) content of ether extract and ADL, and a slightly higher content of crude protein (+0.5% DM;
p > 0.05;
Table 3). Overall, these differences between the two silages remained almost constant for the whole ensiling period, showing a slight reduction at D90. Ash, crude fiber and its fractions (NDF, ADF and ADL) were not influenced (
p > 0.05) by the silage formulation; their values fluctuated slightly during the silage period without significant differences (
p > 0.05;
Table 3).
3.2. Fermentation Characteristics
The fermentation characteristics are reported in
Figure 1 and
Figure 2, and
Table 4. At day 0, SIL-B showed a higher pH value (5.78) than SIL-A (4.84). Thereafter, in both the silages, it decreased (
p < 0.01) with ensiling time, reaching the same value at D7 (4.3), and 3.73 and 3.92, respectively, in SIL-B and SIL-A, at D30, which can be considered the end of lactic fermentation and the silages’ maturation time. At D90, it increased slightly, reaching 4.3–4.4 (
Figure 1). Thus, 30 days can represent the end of lactic fermentation and be considered as the silages’ maturation time, in line with other studies [
36,
44].
The pH value is a pivotal parameter for the proper ensiling process and low pH values could be attributed to the fermentation of water-soluble carbohydrates, by lactic acid bacteria, into organic acids [
45,
46]. In this study, for both the experimental silages, the rapid drop in the pH observed in the first ensiling step and the final pH values near 4.2, which is generally considered as the maximum threshold for well-persevered silages, may be considered indexes of good fermentation quality, according to other authors [
47,
48]. The reduction of pH during the ensiling process is of crucial importance in preventing the growth of undesirable microbes such as clostridia, Enterobacteriaceae, listeria and molds [
45].
The Flieg’s score, based on the pH and dry matter content, is commonly used as an index to classify the quality of silage. In this study, the Flieg’s score recorded at D30 was greater than 100 in SIL-A and SIL-B and, although it showed a decrease at D90, remained above 100 (
Figure 2), indicating that both silage formulations were of a very high quality and stability [
41,
49].
The concentrations of lactic acid and volatile fatty acids and their changes during fermentation are reported in
Table 4.
For both the silages, a rapid increase in lactic acid concentration occurred from the day of ensiling (D0) to D7 after ensiling to reach approximately 20 g kg
−1 DM at D30 (
p < 0.01). At this time, no difference (
p > 0.05) was found between SIL-A and SIL-B for lactic acid concentration. Subsequently, it remained substantially unchanged until D90. In both the experimental silages, the percentage of lactic acid content with respect to the total acids (88.41% and 96.35%, SIL-A and SIL-B, respectively) showed that lactic acid represented the principal compound derived presumably from good fermentative processes leading to high quality silages. In fact, as reported by Ward and Ondzarda [
47], in a homo-fermentative ensiling process, lactic acid should be the principal end-product of fermentation, and the level of its concentration is considered to be a good indicator of silage quality [
50]. Lactic acid (pKa of 3.86) is the fermentation compound that most contributes to the decrease in the pH of silage, because it is about 10 to 12 times stronger than any of the other major organic acids such as acetic (pKa of 4.75) and propionic acid (pKa of 4.87) [
48,
51]. A low pH derived from high lactic acid concentration is considered important for silage preservation [
52]. Moreover, great lactic acid content typically corresponds to low dry matter loss [
47]. In ruminants under normal feeding conditions, lactic acid ingested from silage is converted to propionic acid by rumen microbes such as
Selenomonas ruminantium,
Megasphaera elsdenii or Propionibacteria [
48,
53]. The propionic acid is absorbed by the rumen and is transformed into glucose by the liver. However, very high concentrations of lactic acid, as well as of total acids, in silage have negative effects, depressing feed intake and potentially contributing to subacute acidosis [
53].
The concentration of acetic acid was significantly (
p < 0.01) influenced by the silage formulations and ensiling time, and their interaction (
p < 0.05). As a whole, SIL-A showed higher levels of acetic acid, reaching a concentration of 0.78 g kg
−1 DM compared to 0.23 g kg
−1 DM of SIL-B (
p < 0.01); these values remained almost constant at D90. The concentration of acetic acid in SIL-A was lower compared with the results obtained by Belém et al. [
54]. This is related to the effects of different quantities of grape pomace added to
Calotropis procera silage, where acetic acid concentration increased when grape pomace was increased up to 22%, associated with enterobacteria and heterofermentative lactic acid bacteria activity [
54]. Acetic acid is normally the acid with the second-highest concentration in silage (1 to 3% of DM) [
48]. Its moderate concentration in silage, as for both the silage formulations in this study, is deemed useful for inhibiting yeast and to improve its stability in air. In ruminants, acetic acid from silage can be adsorbed by the rumen and used for energy or for milk or body fat synthesis [
48].
Propionic acid was not detected in SIL-A nor in SIL-B. This could be attributed to appropriate moisture levels in the silages (DM range: 46–59%;
Table 3), according to Kung et al. [
48]. The latter reported that propionic acid may be undetectable in silage with a DM greater than 35%, as in this study, whereas it is commonly present in very wet silage (<25% DM).
Isobutyric acid was affected by silage formulation (
p < 0.01) and significant was the interaction between treatment and time (
p < 0.05;
Table 4). Isobutyric acid concentrations were very low in both the silages, although it was more prevalent in SIL-A, where it ranged from 0.20 at D7 to 0.27 g kg
−1 DM at D90, than in SIL-B (0.02 and 0.05 g kg
−1 DM, at D30 and D90, respectively). The concentration of butyric acid was influenced (
p < 0.05) by the silage formulation, time of ensiling and their interaction (
p < 0.01) (
Table 4). Butyric acid increased with the ensiling period (D7–D90;
p < 0.05), showing, however, low concentrations in the two experimental silages. At D30, its concentration was 0.47 and 0.51 g kg
−1 of DM, respectively, in SIL-A and SIL-B (
p > 0.05), and increased slightly at D90 (
p > 0.05). These values indicated that the two silage formulations had good fermentation quality, below the maximum acceptable concentration of 2.0 g/kg DM for a good quality silage [
55]. Butyric acid formation during fermentation is, in fact, undesirable; the acid is produced due to an increase in
Clostridium spp. if silage has a high humidity or low water-soluble carbohydrates and the pH is too high. In these conditions, Clostridia will increase their numbers and actively convert organic substances and lactic acid into butyric acid, acetic acid, NH
3, CO
2 and H
2, leading to the catabolism of amino acids and amides due to the coupled oxidation of two amino acids (Stickand reactions) [
56]. It has been reported that the fermentation of glucose and lactic acid into butyric acid, CO
2 and hydrogen causes 51.1% of material losses [
57]. In ruminants, butyric acid from silage is metabolized to ketone bodies [
58], but butyric acid, ammonia and amines have been linked with reduced ad libitum feed intake [
59].
The total VFA, comprising acetic, propionic, isobutyric and butyric acids, and its proportion to total acids, was lower (
p < 0.01) in SIL-B compared to SIL-A. However, for both the silages the percentages of VFA were low overall (3.63 and 7.03%, SIL-B and SIL-A;
Table 4), far below the threshold value indicated by Santoso et al. [
46] (20%). This indicates a good fermentation efficiency, as the production of VFA has been related to an inefficient or secondary fermentation of lactic acid [
46].
3.3. Experiment 2—Chemical Composition
The chemical composition of the experimental silages was not influenced (
p > 0.05) by the formulation of by-products nor by the ensiling time. Therefore, it was referred to with the values observed at the ensiling process maturation time (D30;
Table 5); the data at the different ensiling periods were presented in graphic form (
Figure 3), and the discussion is limited to the highly significant findings.
Concerning the chemical composition, DM is considered an important index for nutrition retention after ensiling, and the protein content represents one of the most important attributes of its nutritive value [
60]. Although several authors have reported that DM was not influenced by storage [
61,
62], normally the ensiling process determines losses of DM, reducing the nutritional value of the product as animal feed. These variations have been related to the initial DM content of the ingredients at the time of ensiling and to the conditions of its process [
63]. In this study, DM and the chemical composition of all the experimental silages (SIL-C, SIL-D, SIL-E and SIL-F) at D0 remained almost unchanged over ensiling time (
Figure 3)—probably facilitated by the efficiency of laboratory vacuum ensiling—indicating good retention of nutritional compounds. According to Borreani et al. [
63], the maintenance of DM content could be attributed to the homofermentative process of glucose by LAB that produced only lactate, inducing no DM loss. On the contrary, LAB that ferment glucose heterofermentatively produce 1 mol of carbon dioxide per mol of glucose, leading to DM loss.
Moreover, the protein content of the experimental silages did not show substantial variations during the ensiling period. A possible explanation for the retention of the protein content over time could be related, in particular, to the inclusion of GP in all the silage formulations. GP could have inhibited proteolysis, due to deactivation of proteolytic enzymes and/or through the formation of protein–polyphenol complexes [
64], according to other studies on ensiled forage [
65,
66]. Ke et al. [
67] have reported that the use of grape pomace when ensiling alfalfa could improve the utilization of industrial waste as feed but could also effectively inhibit proteolysis and improve N utilization by ruminants. Fermentation quality, nutritive value and in vitro digestibility of mixed silages containing crop straws and tall fescue was improved by the inclusion of alfalfa [
68], confirming the positive association between a poor ingredient (straw) and another plant [
69]. No important losses in nutritional quality and dry matter were observed in silages produced with broccoli and artichoke by-products [
44].
Phenolic compounds have strong antioxidant activity that affects the antioxidant potential of a feed [
70,
71]. Among the experimental silage formulations, the total phenol content at D30 was the highest in SIL-F and lower in SIL-C (0.72 vs. 0.62 mg GAE g
−1 DM;
Table 5); however, the difference was not significant.
The total phenol content was the highest in SIL-F and lower in SIL-C (0.72 vs. 0.62 mg GAE g
−1 DM;
Table 5); however, the difference was not significant (
p > 0.05). In the four silage formulations, the total phenol content did not decrease during the whole ensiling period studied (D0 to D90;
p > 0.05;
Figure 3), showing a good ensiling process in preserving phenolic content over time as well. A decrease in phenol content during the ensiling process was reported in other studies [
72,
73,
74]. According to Esparza et al. [
75], the stability of phenolic compounds depends on its chemical nature as well as on the overall composition of the matrix because the different compounds present enhance or mitigate the degradation processes.
3.4. Fermentation Characteristics
The fermentation characteristics of the silages in experiment 2 are reported in
Figure 4 and
Figure 5, and
Table 6.
Before ensiling (D0), the pH in the different formulations of silages ranged from 4.82 (SIL-F) to 6.95 (SIL-C). Afterward, in all the silages, the pH decreased as the ensiling progressed, reaching 3.60 to 4.00 (
p > 0.05) at maturation time (D30) and showing a slight increase (range: 3.95–4.30) at D90 (
Figure 4).
Overall, it should be considered that, at the completion of the maturation period (D30), all the silages based on the different combinations of the by-products in the study (experiment 1 and experiment 2) showed pH values below 4.2. This is regarded as an indicator of good fermentation quality in silage [
48], which can prevent the proliferation of enterobacteria. In addition, at D90, the pH can be considered suitable for all the silage formulations, according to Kim et al. [
76]. The latter reported that, in animal feeding, a pH of 4 to 5 of the fermented feed ingredients is suitable because a pH below 4 decreases voluntary feed intake, whereas a pH above 5 may favor microbial spoilage.
At the maturation time (D30), the Flieg’s score of the four experimental silages ranged from 146 to 162. At D90, the Flieg’s score showed a slight decrease, but always remained above the score of 100 (
Figure 5), indicating that all the experimental by-product-based silages were high in quality [
41]. Considering the Fleig’s score of both the experiments, we can confirm the suitability of ensiling by mixing GP, OMWW, CW and straw in terms of nutritional composition and conservation over time.
As observed in experiment 1, in experiment 2, lactic acid was the dominant fermentation product in all the experimental silages produced during the ensiling period (
Table 6). Its concentration was affected by the silage composition (
p < 0.01), time of ensiling and their interaction (
p < 0.05). From D0, when it was not detected, the lactic acid content increased rapidly (
p < 0.01), reaching the maximum level at the end of fermentation (D30) and then remaining stable until D90. At D30, the silages produced with the inclusion of the higher amount of CW (35%; SIL-E and SIL-F) had a higher content of lactic acid, proving significant (
p < 0.05) in comparison with SIL-C (CW at 20%). High lactic acid concentration in silage is considered beneficial for ruminants because it can be metabolized into propionic acid by
Megasphaera elsdenii and used as a precursor for gluconeogenesis [
77].
Acetic acid content was influenced by the by-product combination (
p < 0.01), time of ensiling (
p < 0.05) and their interaction (
p < 0.01). It showed the highest (
p < 0.05) values in SIL-E and SIL-F compared with SIL-C and SIL-D. These differences could be attributed to the additional supply of LAB from cheese whey (+15%) in SIL-E and to the bacteria given through the commercial inoculum (
Lactiplantibacillus plantarum) and/or to an improvement of some environmental factors on which lactic and acetic acid strains could have been depended in SIL-F. It has been reported that acetic acid together with other VFA, by inhibiting the growth of yeasts and mold, improve the aerobic stability of ensiled forage [
78,
79]. A moderate concentration of acetic acid in ruminants’ feed may have positive effects as it can be used to produce energy or fat for body reserves or milk production [
80].
Considering both the main fermentation products: lactic and acetic acids, a ratio of 3:1 has been indicated as the best for a good silage quality [
48]. In this study, this ratio was respected in both the experiments (experiment 1 and experiment 2), from day 7 to day 90.
At the end of silage maturation (D30), propionic acid was found in a negligible amount in all the silages, reaching the maximum concentration of 0.08 g kg
−1 DM in SIL-E (
Table 6). Low propionic acid content might be considered an indicator of well fermented silages. In fact, although Propionibacteria produce this acid from glucose and lactic acid, they are very intolerant of a low pH, and high levels of propionic acid (>0.3 to 0.5%) have been associated with poorly fermented silages and/or the presence of some strains of Clostridia [
53].
Isobutyric and butyric acid contents were affected by silage composition and time of ensiling (
p < 0.05); the ra was significant for the isobutyric acid (
p < 0.05). For both the acids, the contents were the lowest (
p < 0.05) in SIL-D (1.35 g kg
−1 DM) in comparison with the other silages and remained stable at D90 (
Table 6). However, their content was low and very close to the acceptable value of 2.0 g kg
−1 DM, as reported by Moselhy et al. [
81], in SIL-C, SIL-E and SIL-F too. A low butyric acid content has been related to high lactic bacteria populations that reduce enterobacteria, clostridia and yeast populations [
45].
The concentration of total VFA and the proportion of VFA to total acids were the lowest (
p < 0.01) in SIL-D, indicating greater fermentation efficiency compared to the other silages. However, the values found in these silages (SIL-C, SIL-E and SIL-F) were close to 20% (range: 20.6% to 25.8%), which, according to Santoso et al. [
46], is the ideal maximum value. In a study of rice crop residue-based silage, the proportions of VFA to total acids ranged from 31.9% to 37.2%.