Maize processing methods

The maize fields were located at Animal Science Department of Universidade Federal de Viçosa (20°46′ 00″ S and 42°51′28″ W, 649 m above of sea level) in Viçosa, Minas Gerais, Brazil. From October to December 2017, the growing season of maize was characterized as spring with minimum temperatures between 13.8 and 21.5°C, maximum temperatures varying from 22.6 to 34.6°C and precipitations between 0.2 and 87.2 mm.

The maize hybrid, LG 6030 PRO 2 from LG Sementes (Curitiba, SP, Brazil) was planted in sufficient quantity to produce the dry ground maize grain (DMG), RMG, HMM, SNAP, stalklage and whole-plant maize silage for the entire experiment. A dose of 300 kg/ha of N-P-K fertilizer (the fertilizer composed of 8% N, 28% phosphorus and 16% potassium) was applied during the sowing. Maize seeds were planted with a row spacing of 90 cm and on-row plant spacing of approximately 18 cm. Sowing was performed with four-row pneumatic sowing machine. A dose of 100 kg/ha of N (approximately 222 kg/ha of urea) was applied after sowing in two equal aliquots: the first one was provided when the plants had four true leaves and the second aliquot when the plants had eight true leaves.

The DMG and RMG were harvested at the moisture content of 19% and was dried until it reached 13% of moisture. Approximately 6000 kg of dry maize grain was ground in a hammer mill (DMP-2, Nogueiras, São João da Boa Vista, São Paulo, Brazil) using a 3-mm sieve and the DM content was measured (method 934.01; AOAC, 2012). Subsequently, 3000 kg of the ground grain was reconstituted with water to reach a moisture content of 35%, using the following equation below:

$${\rm Amount}\;{\rm of}\;{\rm water\;}( {{\rm L}/{\rm kg\;of}\;{\rm grain}\;{\rm as}\;{\rm fed}} ) = {\rm \;}\displaystyle{{{\rm grain\;DM\;}( {\rm \% } ) -{\rm \;target\;DM\;}( {\rm \% } ) } \over {{\rm target\;DM\;}( {\rm \% } ) }}$$

The grain was soaked, mixed in a cement mixer (Rental Mixer 400L, Menegotti Indústrias Metalúrgicas Ltda., Santa Catarina, Brazil) for 5 min and ensiled with a mean density of 1000 kg of fresh material/m³ to obtain RMG. The remaining 3000 kg of DMG was stored dry in a grain bin.

The HMM and SNAP were harvested when the maize grain contained approximately 40% moisture. Immediately after the harvest, approximately 3000 kg of maize grain was ground in a hammer mill (DMP-2, Nogueiras, São João da Boa Vista, São Paulo, Brazil). Grains were broken into four to six pieces to achieve less than 5% whole kernels and 20% of fines (Hicks and Lake, Reference Hicks and Lake2006; Lardy and Anderson, Reference Lardy and Anderson2016b; Salvo et al., Reference Salvo, Gritti, Daniel, Martins, Lopes, Santos and Nussio2020) and ensiled with a mean density of 1000 kg of fresh feed/m³. Moreover, about 6000 kg of whole ear maize (containing grain, cobs and husks) were chopped in a hammer and knife mill (DPM-JÚNIOR, Nogueiras, São João da Boa Vista, São Paulo, Brazil) and ensiled with a mean density of 600 kg of fresh feed/m³ to obtain the SNAP. The grinder was set to break grains and cobs into small pieces and reduce the husk particle size (Mahanna, Reference Mahanna2008; Akins and Shaver, Reference Akins and Shaver2014; Lardy and Anderson, Reference Lardy and Anderson2016a).

Stalklage was produced with the remaining residue after harvest of SNAP. The material for stalklage and whole-plant maize silage production was harvested with approximately 45 and 30% of DM, respectively. A forage harvester (JF 1600 AT, JF Máquinas Agrícolas, São Paulo, Brazil) set at 22.3 mm theoretical cut length was used (Kononoff et al., Reference Kononoff, Heinrichs and Lehman2003; Cook et al., Reference Cook, Bender, Shinners and Combs2016). Stalklage and whole-plant maize silage were ensiled with a mean density of 400 kg of fresh feed/m³ and 500 kg of fresh feed/m³, respectively.

Penn State Particle Separator (19-, 8-, 4-, and 1.18-mm sieves plus bottom pan; Heinrichs, Reference Heinrichs2013) was used to verify the particle size distribution of all maize corn processing methods. Thus, the grinder or forage harvester was adjusted when needed. The RMG, HMM, SNAP, stalklage and whole-plant maize silage were ensiled in round reinforced concrete pipe silos (1.0 m inside diameter × 1.0 m long × 8.0 cm wall thickness) at the same time, approximately 90 days before the beginning of the experiment.

Animals, facilities and experimental design

Five ruminally cannulated Nellore bulls (age = 8 ± 1.0 mo; initial BW = 265 ± 18.2 kg) were distributed in a 5 × 5 Latin square design. Initially, the animals were individually identified with ear tags, treated for the elimination of internal and external parasites and, housed in a concrete floor tie-stall barn that was equipped with water and feed troughs. The experimental periods were divided into two subperiods of 12 days (Richards et al., Reference Richards, Branco, Bohnert, Huntington, Macari and Harmon2002; Machado et al., Reference Machado, Detmann, Mantovani, Valadares Filho, Bento, Marcondes and Assunção2016; Petzel et al., Reference Petzel, Titgemeyer, Smart, Hales, Foote, Acharya, Bailey, Held and Brake2019) for dietary adaptation and 5 days for in vivo digestibility and in situ degradability technique.

Experimental diets

Five experimental diets were used in the experiment where three of them were composed of 0.30 whole-plant maize silage, 0.10 mineral and protein supplement, and 0.60 (DM basis) of one of the following processing methods: DMG; high-moisture maize (HMM); RMG. The other two diets were composed of 0.10 mineral and protein supplement, 0.80 SNAP and 0.10 stalklage (DM basis; SNAP-80); or 0.10 mineral and protein supplement and 0.90 SNAP (DM basis; SNAP-90). The proportions of SNAP used in this experiment were approximately 0.78 grain, 0.13 cob and 0.09 husk.

The diets were formulated according to BR-CORTE recommendations (Valadares Filho et al., Reference Valadares Filho, Costa e Silva, Gionbelli, Rotta, Marcondes, Chizzotti and Prados2016) to provide, approximately, 125 g CP/kg on a diet DM basis and to support an average daily gain of 1.2 kg/day. Ingredient and chemical composition of the experimental diets are presented in Table 1.

Table 1. Feedstuffs and chemical composition of experimental diets

^a Premix guarantees (per kg of dry matter): 200–220 g of Ca, 10 mg of Co (Min), 500 mg of Cu (Min), 22 g of S (Min), 333 mg of Fe (Min), 178.41 mg of F (Max), 10 g of P (Min), 25 mg of I (Min), 17 g of Mg (Min), 1500 mg of Mn (Min), 1100 mg of monensin, 100 × 109 CFU of Saccharomyces cerevisiae (Min), 6.6 mg of Se (Min), 50 g of Na (Min), 100 000 IU of vitamin A (Min), 13 000 IU of vitamin D3 (Min), 150 IU of vitamin E (Min) and 2000 mg of Zn (Min).

^b Urea + ammonium sulfate in a 9 : 1 ratio.

^c Total digestible nutrients (TDN) was estimated using concentrations of crude protein, ether extract, ash, neutral detergent fibre, lignin, acid and neutral insoluble crude protein as described by BR-CORTE (Valadares Filho et al., Reference Valadares Filho, Costa e Silva, Gionbelli, Rotta, Marcondes, Chizzotti and Prados2016) and then metabolizable energy (ME) was calculated using the following the equations: DE = TDN × 4.4; and ME = 0.9455 × ED – 0.3032 multiplied by a conversion factor of 4.184 to convert Mcal/kg to MJ/kg.

^d Corrected for residual ash and nitrogen compounds.

Evaluation of in vivo digestibility

The diet ingredients were weighed separately, then mixed at the time of feeding, such that a total mixed ration was provided twice per day (08.00 h and 16.00 h). Feed bunks were evaluated each day to quantify refusals and to adjust daily feed allowance to a maximum of 5% of refusals. Feed intake was recorded during the data collection period (from day 13 to 17). Also, diet ingredients and refusals were sampled daily during the collection period and stored at −20°C.

Each diet ingredient and refusals samples were partially dried in a forced-air oven at 55°C for 72 h, ground in a knife mill (Tecnal, Piracicaba, São Paulo, Brazil) using a 1-mm sieve and packed in plastic bags for further laboratory analyses. Each diet ingredient sample was grouped for each period and refusals samples were grouped per animal for each period. These ingredients were analysed individually and used to calculate dietary composition.

From day 15 to 17 of the experimental period, 24 h faecal output was determined for all bulls. Faeces were collected from the concrete floor and placed in 30 l buckets. At the end of each collection day (24 h), the buckets containing the samples were weighed, homogenized and a subsample was collected, dried in a forced-air oven at 55°C for 72 h and ground in a knife mill (Tecnal, Piracicaba, São Paulo, Brazil) with a 1-mm sieve. Furthermore, faecal samples from the 3 day of the collection were combined proportionately for each animal per experimental period according to the dry weight of each collection day and stored in plastic bags for further laboratory analyses.

Evaluation of in situ degradation

From day 13 to 17 of each experimental period, the in situ degradabilities of DMG, HMM, RMG, SNAP-80 and SNAP-90 diets were evaluated. All dietary ingredients were sampled from day 1 to 7 of each experimental period and partially dried in a forced-air oven at 55°C for 72 h. A composite sample from 7 day of the collection was made for each ingredient per period and ground in a knife mill (Tecnal, Piracicaba, São Paulo, Brazil) with a 2-mm sieve.

Approximately 5 g of dried total mixed ration sample was individually weighed into nylon bags (Sefar Nitex; Sefar, Thal, Switzerland; porosity of 50 μm, 8 × 15 cm²) and incubated in each animal (Benedeti et al., Reference Benedeti, Valadares Filho, Zanetti, Silva, Silva, Alhadas, Pereira, Pacheco, Pucetti, Menezes, Silva, Godoi and Santos2019; Menezes et al., Reference Menezes, Valadares Filho, Pacheco, Pucetti, Pereira, Rotta, Zanetti, Silva, Costa e Silva, Detmann, Neville and Caton2019; Silva et al., Reference Silva, Pacheco, Godoi, Silva, Zanetti, Menezes, Pucetti, Santos, Paulino and Valadares Filho2020). To compose the total mixed ration, whole-plant maize silage, DMG, HMM, RMG, SNAP, stalklage, and mineral and protein supplement were weighed separately, maintaining diets composition (DM basis). The incubation for each diet was carried out in the same animal that was receiving the corresponding treatment in the in vivo procedure. Incubation was performed to allow the following ruminal degradation times: 0, 2, 4, 6, 12, 24, 48, 72 and 96 h. The number of bags varied as a function of the incubation time to guarantee enough residual samples after incubation (i.e. more bags per sample were incubated for the longer incubation times relative to the shorter incubation times).

The step by step of rumen in situ incubation procedure is presented in Supplementary Material. The bags containing previously weighed diet samples were attached to a steel chain (90 × 2 cm²; Menezes et al., Reference Menezes, Valadares Filho, Pacheco, Pucetti, Pereira, Rotta, Zanetti, Silva, Costa e Silva, Detmann, Neville and Caton2019) with a weight (300 grams) at the end to allow for complete immersion within ruminal contents. The steel chain was placed into the rumen and attached to the cannula cap using a plastic rope (15 cm) fixed to a PVC hose pipe (3 cm). The cannula was closed with the PVC hose pipe outside. The bags were placed into the rumen in reverse order so that all bags were removed at the same time. After the incubation period, the bags were washed in running water until the rinse water was clear (Silva et al., Reference Silva, Pacheco, Godoi, Silva, Zanetti, Menezes, Pucetti, Santos, Paulino and Valadares Filho2020). The 0 h bags were not incubated in the rumen but were rinsed using the same procedure as incubated bags. The nylon bags with samples were oven-dried at 55°C for 72 h. In sequence, bags were placed in an oven at 105°C for 2 h and weighed. The residues of each diet were removed from the nylon bags, ground in a knife mill (Tecnal, Piracicaba, São Paulo, Brazil) with a 1-mm sieve, stored in a labelled plastic bag for further chemical composition.

Chemical analyses

Diet ingredients samples were analysed for DM, OM, N and ether extract (EE), according to AOAC (2012) method numbers 934.01, 930.05 and 981.10 and AOAC (2006) method number 945.16, respectively. The neutral detergent fibre (NDF) analysis was performed according to techniques described by Mertens (Reference Mertens2002), without the addition of sodium sulfite, but with the addition of thermostable alpha-amylase to the neutral detergent. The NDF content was corrected for residual ash and protein (apNDF). Estimations of neutral detergent insoluble nitrogen (NDIN) followed the technique described by Licitra et al. (Reference Licitra, Hernandez and Van Soest1996). The starch analysis was performed following the recommendations of Silva et al. (Reference Silva, Godoi, Valadares Filho, Zanetti, Benedeti and Detmann2019). Refusals, faeces and in situ residues were analysed for DM, OM and starch contents, according to previously described methods.

Degradation models and partition of starch digestion

For the in situ evaluation, the DM, OM and starch degradation profiles were estimated using the Ørskov and Mcdonald (Reference Ørskov and McDonald1979) asymptotic function:

$$Y_t = {\rm \;}a{\rm \;} + {\rm \;}b{\rm \;} \times ( {1-{\rm \;}{\rm e}^{( {-ct} ) }} ) $$

where Y_t = degraded fraction of DM, OM or starch at time ‘t’, g/kg; a = readily soluble fraction, g/kg; b = potentially degradable fraction in the rumen, g/kg; c = rate constant for degradation of b, per hour; t = time, hour.

The starch effective degradability (ED) for each diet was calculated according to the equation of Ørskov and McDonald (Reference Ørskov and McDonald1979) and corrected for 6% of starch washing out of the bag that would escape rumen degradation (Offner and Sauvant, Reference Offner and Sauvant2004):

$$\eqalign{{\rm E}{\rm D}_{{\rm measured}} = \,& ( {a + ( {b \times c} ) / ( {c + kp} ) } ) \cr {\rm E}{\rm D}_{{\rm corr}} = \,& {\rm E}{\rm D}_{{\rm measured}}\ndash ( {0.06 \times a} ) } $$

where ED = starch effective degradability, g/kg; a = readily soluble fraction, g/kg; b = potentially degradable fraction in the rumen, g/kg; c = rate constant for degradation of b, per hour; kp = passage rate of starch. The kp for each diet was obtained by in vivo ruminal emptying procedure and it can be found in Godoi et al. (Reference Godoi, Silva, Silva, Pucetti, Pacheco, Souza, Lage, Rennó, Schoonmaker and Valadares Filho2021).

The ruminal digestibility of starch (Rd) was predicted according to the equation proposed by Offner and Sauvant (Reference Offner and Sauvant2004):

$${\rm Rd\ } = 0.302 + 0.59{\rm ED}$$

The intestinal digestibility of starch (Id) was obtained by the difference between the total digestibility observed from in vivo procedure and the prediction of ruminal digestibility as described above.

$${\rm Id\ } = {\rm Total}\;{\rm tract\ }-{\rm Rd}$$

Use of in situ techniques to predict in vivo digestibility

Silva et al. (Reference Silva, Pacheco, Godoi, Silva, Zanetti, Menezes, Pucetti, Santos, Paulino and Valadares Filho2020) performed simultaneous in situ and in vivo trials and suggested that the optimal in situ incubation time required to accurately estimate in vivo digestibility of DM, OM and starch was 24 h. Moreover, a multi-study analysis was performed by Benedeti et al. (Reference Benedeti, Valadares Filho, Zanetti, Silva, Silva, Alhadas, Pereira, Pacheco, Pucetti, Menezes, Silva, Godoi and Santos2019) to develop and validate an equation to estimate in vivo OM digestibility from in situ methods. According to these authors, stepwise regression results showed that c contributed significantly to predictions of in vivo digestibility. Thus, the equation below was suggested to estimate the in vivo digestibility coefficient of OM using the in situ technique:

$$Y = 0.5695204 + 2.8597612 \times c$$

where Y = in vivo digestibility coefficient of OM; c = rate constant for degradation of b estimated using different in situ incubation points, per hour; t = time, hour.

The in vivo digestibilities coefficients of DM, OM and starch estimated for each animal were compared to the in situ estimations using a single 24 h in situ incubation point suggested by Silva et al. (Reference Silva, Pacheco, Godoi, Silva, Zanetti, Menezes, Pucetti, Santos, Paulino and Valadares Filho2020) and the equation proposed by Benedeti et al. (Reference Benedeti, Valadares Filho, Zanetti, Silva, Silva, Alhadas, Pereira, Pacheco, Pucetti, Menezes, Silva, Godoi and Santos2019). It is worth mentioning that the present study adopted a similar in vivo and in situ procedures to the studies mentioned above.

Statistical analysis

Statistical analysis for in situ degradation parameters and partition of starch digestion was performed using the MIXED procedure of SAS 9.4 (SAS Institute Inc., Cary, NC, USA). Data were analysed using the following model:

$$Y_{ijk} = \mu + D_i + a_j + p_k + e_{ijk}$$

where Y_ijk = response variable; μ = overall mean; D_i = fixed-effect of ith dietary treatment (5 levels); a_j = random effect of the jth animal (5 levels); p_k = random effect of the kth period (5 levels) and e_ijk = residual error, assuming e_ijk ~ N (0, s ²).

Least-squares means were separated using Fisher's least significant difference test. Results were deemed significant when P ≤ 0.05.

The estimated DM, OM and starch digestibility (n = 25) using a single 24 h in situ incubation point suggested by Silva et al. (Reference Silva, Pacheco, Godoi, Silva, Zanetti, Menezes, Pucetti, Santos, Paulino and Valadares Filho2020) and the OM digestibility equation proposed by Benedeti et al. (Reference Benedeti, Valadares Filho, Zanetti, Silva, Silva, Alhadas, Pereira, Pacheco, Pucetti, Menezes, Silva, Godoi and Santos2019) were compared to the values observed in the in vivo trial (n = 25) using the following regression model:

$$Y = \beta _0 + \beta _1 \times X, \;$$

where X = predicted in situ digestibility; Y = observed values using the in vivo trial; β ₀ = intercept of equation; and β ₁ = slope of equation. Regression was evaluated according to the following statistical hypotheses (Mayer et al., Reference Mayer, Stuart and Swain1994):

$$H_0\colon \beta _0 = 0\;{\rm and}\;\beta _1 = 1, \;\;{\rm and}\;H_a\colon {\rm not\ }H_0$$

If the null hypothesis was not rejected, it could be concluded that the equations and the 24 h of in situ incubation time accurately estimate the apparent digestibility of DM, OM and starch. Slope and intercept were separately evaluated to observe where equations have possible errors. Estimates were evaluated using the estimated value of mean square error of prediction (MSEP) and its components (Bibby and Toutenburg, Reference Bibby and Toutenburg1977):

$$\eqalign{{\rm MSEP} = \,& {\rm SB\ } + {\rm MaF\ } + {\rm MoF\ } = 1 / n \sum _i = _1( {X_i- Y_i} ) ^2, \;\cr {\rm SB } =\, & ( {X - Y} ) ^2, \;\cr {\rm MaF} =\, & ( {s_X\ndash s_Y} ) ^2, \;\cr {\rm MoF} =\, & 2s_Xs_Y( {1 \ndash R} ) , \;} $$

where X = predicted values; Y = observed values; MSEP = mean squared error of prediction; SB = squared bias; MaF = component relative to the magnitude of random fluctuation; MoF = component relative to the model of random fluctuation; s_X and s_Y = standard deviations of predicted and observed values, respectively; and R = Pearson linear correlation between predicted and observed values.

For all variance and covariance calculations, the total number of observations was used as a divisor since it was a prediction error estimate (Kobayashi and Salam, Reference Kobayashi and Salam2000). Prediction of efficiency was determined by estimating the correlation and concordance coefficient (CCC) or reproducibility index described by Tedeschi (Reference Tedeschi2006). Validation analyses were performed with the Model Evaluation System [MES; version 3.1.16 (Tedeschi, Reference Tedeschi2006)] and significance was established at α = 0.05.