J. Dairy Sci.

101:6047–6054
https://doi.org/10.3168/jds.2017-13687
© American Dairy Science Association®, 2018.

Effect of re-ensiling on the quality of sorghum silage

G. V. S. dos Anjos,*1 L. C. Gonçalves,* J. A. S. Rodrigues,† K. M. Keller,‡ M. M. Coelho,* P. H. F. Michel,*
D. Ottoni,* and D. G. Jayme*
*Department of Animal Science, Veterinary School, Federal University of Minas Gerais, Minas Gerais, 31270-901, Brazil
†Brazilian Agricultural Research Corporation, EMBRAPA—Maize and Sorghum, Sete Lagoas, Minas Gerais, 35701-970, Brazil
‡Department of Preventive Veterinary Medicine, Veterinary School, Federal University of Minas Gerais, Minas Gerais, 31270-901, Brazil

ABSTRACT pH and loss of dry matter of the silages (4.23 vs. 3.98
and 14.05 vs. 7.82%, respectively) and therefore did not
The commercialization of silage in many countries, provide any benefits in this study.
including Brazil, has increased in recent years. Re- Key words: aerobic stability, Propionibacterium
ensiling of previously ensiled forage occurs when silage acidipropionici, Lactobacillus plantarum, relocation,
is relocated from one farm to another, where it will be silage quality
compacted and sealed again. During this process, silage
is exposed to oxygen before being ensiled, which may
INTRODUCTION
affect its quality. We exposed sorghum silage to air dur-
ing the anaerobic storage phase to simulate the trans- Forage conservation in the form of silage occurs under
portation of silages between farms. Experimental treat- anaerobic conditions, where soluble carbohydrates are
ments included silage exposed to air for 0 or 12 h, with fermented to organic acids that reduce the pH of the
or without the use of an inoculant containing a mixture medium, thus conserving the material by inhibiting the
of Lactobacillus plantarum and the propionic bacteria growth of undesirable microorganisms. When exposed
Propionibacterium acidipropionici (1 × 106 cfu/g of for- to air, the silage may deteriorate and lose its nutritional
age; Biomax corn, Lallemand, Saint-Simon, France), value because of growth of detrimental microorganisms
totaling 4 treatments: conventional silage, conventional such as yeasts (Pahlow et al., 2003).
silage with inoculant use, re-ensilage after exposure to In Brazil, the commercialization of silage has become
air, and re-ensilage after exposure to air with use of an more frequent owing to reduced crop yields as a result
inoculant. The sorghum was stored in experimental si- of climatic variations during the rainy season and oc-
los containing about 9.0 kg of fresh forage per replicate. curs mainly on small farms that do not have sufficient
Treatments were tested in a factorial 2 × 2 design with manpower and machinery for the production of silage.
5 replicates each. Chemical composition, in vitro dry Additional factors associated with the decision to pur-
matter digestibility, fermentative characteristics, losses chase silage from other farms include the excessive losses
(due to gas, effluents, and total dry matter), microor- during storage and regional topography (de Lima et al.,
ganism counts, and aerobic stability of sorghum silage 2016). To meet the demands of this growing market,
were evaluated. Dry matter content of sorghum before some farmers have started specializing in silage produc-
ensiling was 273.12 g/kg. The 12-h re-ensiling process tion for commercialization. As a result, forage ensiled
increased the effluent loss of the silage when compared on the farm must be unpacked and transported to other
with conventional silage (456.42 vs. 201.19 g/kg of FM, locations, where it will be compacted and sealed again.
respectively). In addition, re-ensiled silages presented During this process, the material is exposed to air for
lower concentrations of lactic acid and higher concen- an indeterminate period, which affects the quality of
trations of propionic acid than the silages that had not the silage (Michel et al., 2017). The existing literature
been opened during storage. The aerobic stability of on the re-ensiling process is very scarce and usually
silage was not affected by the re-ensiling process and does not take into account the characteristics of the
the use of inoculant. The use of inoculant increased the tropical climate. As a result, farmers and technicians
are faced with numerous uncertainties regarding pos-
sible detrimental effects of silage transport.
Silage stability in the presence of oxygen is an impor-
Received August 15, 2017.
Accepted February 21, 2018. tant factor that determines the quality and nutritional
1
Corresponding author: g.dosanjos@hotmail.com value of this material (Filya et al., 2004), and bacterial

6047
6048 DOS ANJOS ET AL.

inoculants may attenuate the decline in silage quality and the propionic bacteria P. acidipropionici MA26/4U
during the transportation process. The aim of this (Biomax corn, Lallemand, Saint-Simon, France). In-
study was to evaluate the effect of re-ensiling and the oculation (1 × 106 cfu/g of forage) was performed at
use of microbial inoculants (Lactobacillus plantarum ensiling. The inoculum was diluted in water and was
and Propionibacterium acidipropionici) on the quality uniformly sprayed onto the forage with a backpack
of sorghum silage. sprayer and mixed.
Fifty-six days after ensiling, half of the silos were
MATERIALS AND METHODS opened and re-ensiled. The silos were opened in a shed,
and the material was removed and re-ensiled after 12 h
Planting, Harvesting, and Ensiling of air exposure. The temperature ranged between 20.0
and 26.0°C during the exposure period, and relative
The BRS 655 sorghum crop was planted in Novem- humidity was between 61 and 92% (data obtained from
ber 2013 in the experimental area of Embrapa Milho the automatic weather station of the Brazilian National
e Sorgo, located in Sete Lagoas, Minas Gerais, Brazil Institute of Meteorology, located 2.4 km from the shed).
(19°28′S, 44°15′W, altitude 732 m). Sorghum was plant- All silos were opened 240 d after the re-ensiling pro-
ed in 5 blocks, each with 2,000 m2, to provide samples cess (296 d of storage in total). Samples were obtained
representing a range of moisture and soil fertility. The for analysis of chemical composition, in vitro digest-
space between planting lines was 70 cm. The crop was ibility, silage quality (pH, NH3-N, and lactic, acetic,
fertilized at planting (32, 112, 64, and 2 kg/ha of N, P, propionic, and butyric acids), loss of material (gas, ef-
K, and Zn, respectively), and a top-dressing of 100 kg fluent, and total DM), aerobic stability, and total count
of N/ha was applied 35 d after planting. of molds, yeasts, and aerobic bacteria.
When the grain reached the milk stage, the crop was
harvested and chopped into 1-to-2-cm segments with a
conventional forage harvester (JF C120 AT; JF Agri- Chemical Analysis and In Vitro DM Digestibility
cultural Machines, Itapira, Brazil). The chopped forage
was sampled (1 sample of 800 g/block) and analyzed Fresh forage samples were oven-dried at 55°C for
before being inoculated. Half of the material harvested 72 h and were subsequently processed in a knife mill
in each block was weighed and inoculated; the other (sieve size = 1 mm; Thomas Wiley model 4, Thomas
half was weighed and water was added at a rate identi- Scientific, Swedesboro, NJ). Dry matter at 105°C, CP,
cal to that of the inoculated silage (200 mL for 100 kg ash, and ether extract (EE) were determined according
of fresh forage). A total of 9.0 ± 0.1 kg of sorghum was to procedures outlined by AOAC International (2005).
placed in each silo, and the forage was then manually Cell wall components (NDF, ADF, and lignin) were
compacted into experimental silos. For each block, a determined by the sequential method according to
silo was made for each of the 4 treatments, amounting methods described by Van Soest et al. (1991). The
to 20 experimental silos in total. Silos were made from NDF and ADF residues were subjected to ash and CP
20-L plastic buckets equipped with a Bunsen valve to analyses to determine the amount of neutral detergent
allow for the release of fermentation gas. A cotton bag insoluble protein and acid detergent insoluble protein.
was placed inside each bucket with approximately 2 kg These values were
​​ used to correct NDF and ADF for
of dry sand to allow the measurement of effluent. ash and protein. The NFC levels were​​ calculated us-
ing the equation proposed by NRC (2001); NFC =
100 – (% NDF + % CP + % EE + MM), where MM
Experimental Design is the amount of mineral matter or ash. In vitro DM
The re-ensiling process and the use of a microbial digestibility (IVDMD) was determined on the DaisyII
inoculant during ensiling and re-ensiling of silage af- digestion apparatus (Ankom Technology, Fairport, NY)
ter 12 h of air exposure were evaluated, and 12 h was according to methods described by Tilley and Terry
considered to be the minimum time required for ensil- (1963) and adapted by Holden (1999). Ruminal fluid
ing and re-ensiling processes in Brazil. The treatments was collected from a cannulated cow that was fed a
were arranged in a 2 × 2 factorial scheme, with 5 repli- diet comprising 10 kg (DM) of sorghum silage and 3 kg
cates each (blocks). The first variable examined was the (DM) of commercial concentrate.
re-ensiling process (with or without), and the second
variable investigated was the use of inoculant (used or Analysis of Fermentative Parameters
not used).
The inoculum comprised the facultative heterofer- Silage juice was extracted by a hydraulic press (2.5
mentative lactic acid bacteria L. plantarum MA18/5U kgf/cm2) to determine the pH, ammoniacal nitrogen

Journal of Dairy Science Vol. 101 No. 7, 2018

 RE-ENSILING SORGHUM SILAGE 6049

(NH3-N), and organic acids. The pH was measured with in a room at 25 ± 1°C to evaluate aerobic stability. Si-
a digital potentiometer (HI 221, Hanna Instruments, lage temperature was monitored every 10 min with the
Woonsocket, RI). Distillation to determine NH3-N was aid of a temperature data logger inserted 15 cm into the
performed in Kjeldahl equipment using magnesium center of mass. In addition, 1.5 kg of silage was placed
oxide and calcium chloride as a neutralizing medium in another set of buckets to track changes in microbial
for the evaporation of ammonia, with boric acid as the and pH counts. Samples from these buckets were taken
receptor solution and 0.1 M hydrochloric acid as the on d 0, 2, 6, and 10 following the opening of silos for
titrant. The levels of organic acids (acetic, lactic, propi- pH evaluation. Aerobic deterioration was considered to
onic, and butyric acid) were determined by GC (GC-17 have occurred if the temperature difference between the
Shimadzu gas chromatograph; Shimadzu Corp., Kyoto, materials and ambience reached 2°C (Ranjit and Kung,
Japan) equipped with a flame ionization detector and 2000). Microorganisms were counted in silage samples
fitted with a Kukol capillary column according to the on d 0 and when they lost aerobic stability or on d 10
methodology described by Playne (1985). The gas chro- in silages that remained stable.
matograph was operated isothermally with a column
temperature of 200°C and an inlet and detector tem-
Microbiological Analyses
perature of 225°C. The pH values of samples subjected
to the aerobic stability test were determined as follows:
Samples were collected for total aerobic microbial
fresh silage (9 g) was added to 60 mL of distilled water,
counts (yeasts, mold, and aerobic bacteria). Analysis of
and pH values were measured after 30 min (Silva and
the microbiota was performed using a standard disper-
Queiroz, 2002).
sion plate method. Total bacterial counts were deter-
mined aerobically on plate count agar (Difco, Sparks,
Analysis of Loss MD) following an incubation period of 1 to 3 d at 36
± 1°C. Total yeast counts were determined on tryptone
The weight of the empty silos plus lid plus dry sand glucose yeast extract agar according to Pitt and Hock-
bags was recorded before the ensiling process. The silos ing (2009). Next, aerobic samples were incubated for 1
were then filled with forage, compacted, covered, sealed to 3 d at 30 ± 1°C. Total mold count was determined
with adhesive tape, and weighed again. on dichloran rose-bengal chloramphenicol agar accord-
Silos subjected to the re-ensiling procedure were ing to Pitt and Hocking (2009) after an aerobic incuba-
opened at 56 d and weighed before and after forage tion period of 5 to 7 d at 25 ± 1°C. The plates were
removal to determine the production of gases and ef- examined daily for typical colony and morphological
fluents. The DM content in the forage was also deter- characteristics associated with each growth medium.
mined. The sand deposited at the bottom of each silo Total microbial counts were expressed as colony form-
was replaced, after which the empty set was weighed. ing units per gram. All microbial counts were log10
The forage was re-ensiled after 12 h of air exposure. transformed to obtain lognormal distribution.
After filling and sealing, the experimental silos were
weighed again to determine the total weight.
We weighed all silos 296 d after the initial ensiling to Statistical Analyses
determine the loss of gas. Silages were then removed,
and the silos were weighed to quantify the effluent pro- The results of chemical composition, in vitro digest-
duced. Total DM loss was estimated as the difference ibility, silage quality, losses, aerobic stability, and total
between the final and initial dry weight of the experi- counts of molds, yeasts, and aerobic bacteria were ana-
mental silos in relation to silage DM weight, minus the lyzed in randomized blocks in a 2 × 2 factorial scheme
weight of the ensilage set before the silos were opened with 5 replicates. Analyses were carried out as follows:
(Jobim et al., 2007). Gas, effluent, and total DM losses
for the re-ensiled silages were obtained from the sum of Yijk = μ + Ri + Ij + Nij + Bk + eijk,
losses during the opening for re-ensiling and the final
opening. where Yijk = observed response values; μ = overall
mean; Ri = the effect of the ith level of re-ensiling time
Aerobic Stability Test (0 or 12 h); Ij = the effect of the jth level of inoculant
(with or without application); Nij = the effect of ith re-
Plastic buckets (23 cm height and 66 cm circumfer- ensiling and jth inoculant interaction; Bk = fixed effect
ence) containing 1.5 kg of silage per replicate were placed of the kth block (1, 2, 3, 4, 5); and eijk = random error.

Journal of Dairy Science Vol. 101 No. 7, 2018

6050 DOS ANJOS ET AL.

The pH data on d 0, 2, 6, and 10 of the aerobic The DM content of the fresh sorghum before ensiling
stability test were analyzed in randomized blocks with and inoculation was 273.12 g/kg; the sorghum con-
subdivided plots. The sources of variation were blocks, tained 66.5 g/kg of DM CP, 42.1 g/kg of DM ash, 23.7
treatments in 2 × 2 factorial arrangement (plots), and g/kg of DM EE, 558.7 g/kg of DM NDF, 301.9 g/kg of
days of evaluation (subplots). Analyses were carried DM ADF, and 47.93 g/kg of DM lignin. The chemical
out as follows: composition of the re-ensilaged silages did not present
alterations regarding the conventional silages (Table 1).
Yijk = μ + Ri + Ij + Dl + Nij + Oil + Pjl Silages exposed to air deteriorate as a result of aerobic
microbial activity, leading to loss of nutritional value
+ Qijl + Bk + eijk, (Filya et al., 2006). However, the deterioration process
depends on the quality of the ensiled material and the
where Yijk = observed response values; μ = overall air exposure time (Chen and Weinberg, 2014; de Lima
mean; Ri = the effect of the ith level of re-ensiling time et al., 2016).
(0 or 12 h); Ij = the effect of the jth level of inoculant Silages inoculated with microbial additives exhibited
(with or without application); Dl = the fixed effect of higher levels of neutral detergent insoluble protein
the lth days of evaluation (0, 2, 6, and 10); Nij = the compared with noninoculated silages (23.3 vs. 20.2 g/
effect of the ith re-ensiling and jth inoculant interac- kg of DM). In addition, NDF values were higher in
tion; Oil = interaction between the ith re-ensiling and inoculated silage materials. Furthermore, inoculated si-
lth days of evaluation; Pjl = interaction between the jth lages demonstrated lower NFC content compared with
inoculant and lth days of evaluation; Qijl = the interac- untreated silages. Filya et al. (2000) also reported lower
tion between the ith re-ensiling time, the jth inoculant NFC content in silages inoculated with microbial addi-
present, and the lth days of evaluation; Bk = fixed effect tives than in noninoculated silages. According to the
of the kth block (1, 2, 3, 4, 5); and eijk = random error. authors, lower NFC contents represent a greater extent
Data were submitted to ANOVA. When significant of silage fermentation.
interactions were identified, further analyses of simple The success of an introduced inoculant depends on
effects were conducted (sliced ANOVA). When interac- several factors, such as the properties of plants and
tions were found to be insignificant, the effects of the inoculants used (Kristensen et al., 2010; Muck, 2010).
re-ensiling process and inoculation were analyzed sepa- It was observed that the use of inoculant led to a de-
rately by F-tests at 5% significance. Regression analysis crease in silage quality (decreased NFC and IVDMD).
at 5% significance was carried out to examine the effect Because NFC are highly digestible organic compounds,
of days of evaluation on silage pH during the aerobic their reduction in concentration may result in reduced
stability test. The analysis was performed using the digestibility of the material (McDonald and Heron,
PROC GLM software from SAS (SAS Institute Inc., 2001), as observed in silages that received microbial in-
Cary, NC). oculants (Table 1). There was no difference in IVDMD
between conventional and re-ensiled silages. The use
of microbial inoculants resulted in a 3% reduction in
RESULTS AND DISCUSSION IVDMD compared with noninoculated silages (616.6
vs. 635.2 g/kg of DM).
The process of re-ensiling has been adopted as a Table 2 outlines measurements of fermentation pa-
common practice for Brazilian farmers in recent years. rameters. Results indicated that re-ensiling reduced the
Once exposed to oxygen, the ensiled material can be NH3-N content of the material. In addition, re-ensiled
stored for several months after re-ensiling. Most farm- materials had lower lactic acid and higher propionic
ers use the re-ensiled material within a short time (ap- acid contents compared with conventional silages (29.0
proximately 60 d or less). However, many farmers pur- vs. 46.7 and 8.7 vs. 2.7 g/kg of DM, respectively). This
chase large amounts of silage at low prices and re-ensile was expected, as exposure of silage to air enables the
the silage for use throughout the year. Thus, aiming growth of deteriorating aerobic microorganisms such as
to simulate scenarios like this, we decided to keep the molds, yeasts, and certain bacteria, which can use VFA
stored re-ensiled material for 240 d. for metabolism. One of the main products consumed by
Our results demonstrated that re-ensiling sorghum yeasts is lactic acid, producing CO2 and water in the
silage did not significantly affect silage quality. In addi- process (Tabacco et al., 2011).
tion, the use of microbial inoculant containing L. plan- The use of inoculant increased silage pH, reduced lac-
tarum and P. acidipropionici did not cause improve- tic acid content of the material, and increased propionic
ments in the silages. acid concentration. It has been shown that in addition

Journal of Dairy Science Vol. 101 No. 7, 2018

 RE-ENSILING SORGHUM SILAGE 6051
Table 1. Chemical composition (g/kg of DM) of sorghum silage treated with inoculants and after re-ensiling

Treatment2

Control Inoculant P-value3

Parameter1 SIL RE   SIL RE SEM I R I×R

DM 263.9 264.3   245.8 259.2 3.70 NS NS NS
Ash 41.8 45.1   46.6 43.7 0.76 NS NS <0.05
CP 83.2 90.2   84.6 91.7 1.67 NS NS NS
NDIP 18.9 21.6   23.9 22.7 0.62 <0.05 NS NS
ADIP 12.9 13.8   13.5 13.1 0.28 NS NS NS
EE 27.7 24.5   25.8 25.4 0.84 NS NS NS
NDF 600.7 619.5   634.0 648.9 6.55 <0.01 NS NS
NDFap 568.4 584.5   613.2 611.7 7.32 <0.01 NS NS
ADF 337.9 334.0   325.4 349.0 5.44 NS NS NS
ADFap 324.3 318.9   310.9 334.8 5.42 NS NS <0.05
NFC 246.7 220.6   202.6 190.3 7.15 <0.01 NS NS
Lignin 40.9 37.9   37.9 38.5 0.98 NS NS NS
IVDMD 636.8 633.5   628.3 604.8 5.23 <0.05 NS NS
1
NDIP = neutral detergent insoluble protein; ADIP = acid detergent insoluble protein; EE = ether extract;
NDFap = neutral detergent insoluble fiber corrected for ash and protein; ADFap = acid detergent insoluble
fiber corrected for ash and protein; IVDMD = in vitro DM digestibility.
2
Experimental treatments included silage exposed to air for 0 or 12 h with or without the use of an inoculant
containing a mixture of Lactobacillus plantarum and the propionic bacteria Propionibacterium acidipropionici
(1 × 106 cfu/g of forage; Biomax corn, Lallemand, Saint-Simon, France), totaling 4 treatments: conventional
silage (SIL), conventional silage with inoculant use, re-ensilage (RE) after exposure to air, and re-ensilage after
exposure to air with inoculant use.
3
I = inoculant effect; R = re-ensiling effect; I × R = interaction effect.

to carbohydrates, P. acidipropionici uses lactic acid Silo fermentation produces volatile compounds that
as substrates during the fermentation process (Filya may lead to DM reduction in the ensiled material
et al., 2004). It is possible that in this study, the use (Kristensen et al., 2010). The re-ensiled silages showed
of inoculant promoted the reduction of lactic acid and higher losses due to effluent compared with conven-
an increase of propionic acid compared with untreated tional silages (456 vs. 210 g/kg, respectively; Table 2).
silages. According to Michel et al. (2017), the higher effluent

Table 2. Fermentation quality parameters (g/kg of DM unless noted) of sorghum silage after re-ensiling and
treatment with inoculants

Treatment2

Control Inoculant P-value3

Parameter1 SIL RE   SIL RE SEM I R I×R

pH 3.98 3.98   4.05 4.42 0.07 <0.01 NS NS
NH3-N/TN (g/kg) 31.5 25.7   34.9 28.0 0.14 NS <0.05 NS
Lactic acid 75.65 42.74   17.82 15.23 7.03 <0.01 <0.05 NS
Acetic acid 11.66 38.45   29.91 30.38 4.60 NS NS NS
Propionic acid 1.46 4.68   3.97 12.78 1.43 <0.05 <0.01 NS
Butyric acid 4.15 1.14   7.20 7.17 1.09 NS NS NS
Gas loss (% of DM) 5.47 6.37   11.08 12.55 1.13 <0.05 NS NS
Effluent loss (g/kg of FM) 184.64 415.97   235.74 496.86 30.51 <0.01 <0.01 NS
Total loss (% of DM) 6.60 9.04   14.21 13.90 1.37 <0.05 NS NS
1
NH3 = ammonia nitrogen; TN = total nitrogen; FM = fresh matter.
2
Experimental treatments included silage exposed to air for 0 or 12 h with or without the use of an inoculant
containing a mixture of Lactobacillus plantarum and the propionic bacteria Propionibacterium acidipropionici
(1 × 106 cfu/g of forage; Biomax corn, Lallemand, Saint-Simon, France), totaling 4 treatments: conventional
silage (SIL), conventional silage with inoculant use, re-ensilage (RE) after exposure to air, and re-ensilage after
exposure to air with inoculant use.
3
I = inoculant effect; R = re-ensiling effect; I × R = interaction effect.

Journal of Dairy Science Vol. 101 No. 7, 2018

6052 DOS ANJOS ET AL.

Table 3. Aerobic stability (h) and microbial total count (log10 cfu/g) under various conditions

Treatment1

Control Inoculant P-value2

Parameter SIL RE   SIL RE SEM I R I×R

Aerobic stability 240 225.6   240 206.4 8.49 NS NS NS
Microbial count at silage opening            
  Aerobic bacteria 5.40 5.82   6.04 5.69 0.15 NS NS NS
 Yeasts 2.89 2.53   3.11 3.09 0.17 NS NS NS
 Molds 2.90 2.54   2.25 3.02 0.18 NS NS NS
Count at loss of stability            
  Aerobic bacteria 5.86 6.92   6.09 5.45 0.18 NS NS NS
 Yeasts 5.76 6.03   1.99 2.15 0.54 <0.05 NS NS
 Molds 4.01 6.31   3.88 2.14 0.43 NS NS NS
1
Experimental treatments included silage exposed to air for 0 or 12 h with or without the use of an inoculant
containing a mixture of Lactobacillus plantarum and the propionic bacteria Propionibacterium acidipropionici
(1 × 106 cfu/g of forage; Biomax corn, Lallemand, Saint-Simon, France), totaling 4 treatments: conventional
silage (SIL), conventional silage with inoculant use, re-ensilage (RE) after exposure to air, and re-ensilage after
exposure to air with inoculant use.
2
I = inoculant effect; R = re-ensiling effect; I × R = interaction effect.

losses in re-ensiled silages may be related to the fact (Muck, 2010), and even some microorganisms that may
that this material undergoes 2 compaction processes develop, as is the case of yeasts, have their limited
because the compaction allows the removal of water growth rate under these conditions (Pitt and Muck,
from the plant cells and thus promotes greater effluent 1993).
production. However, no differences in total DM losses On the other hand, although microorganism counts
were detected between the 2 treatments. Similarly, were similar at silo opening (Table 3), inoculated si-
Chen and Weinberg (2014) also did not detect any dif- lages demonstrated lower yeast count compared with
ference in DM loss of re-ensiled corn silage with up to untreated silages following the stability test (2.07 vs.
48 h of air exposure. Our results indicated that inocu- 5.89 log cfu/g). This may be due to higher production
lated silages showed higher effluent and gas production of propionic acid in inoculated silage, which acts as
losses compared with noninoculated silages (366 vs. 300 an important inhibitor of yeast growth (Filya et al.,
and 11.8 vs. 5.9 g/kg, respectively). Dry matter losses 2004). Tabacco et al. (2009) reported that yeast counts
in silages inoculated with microbial additives were 6 above 5 log10 cfu/g may reduce silage aerobic stability.
percentage units higher compared with noninoculated However, the high yeast counts observed in the present
silages (14.05 vs. 7.82%). Similarly, Tabacco et al. study did not have an effect on silage stability.
(2011) also reported higher DM losses in silages inocu- During the aerobic stability test, an increase in pH
lated with microbial additives compared with those in was observed with increased air exposure time for all
conventional silages. treatments (Figure 1). At silo opening, yeasts that
The aerobic stability of silage was not affected by use lactic acid under aerobic conditions may develop,
the re-ensiling process and the use of inoculant (Table thereby raising silage pH (Muck, 2010). An increase
3). During the course of the stability test (240 h), the in pH values with greater air exposure was observed
temperature of the material remained relatively stable across all treatments. However, this increase was higher
(mean temperature of the silages at the end of the test in silages inoculated with microbial additives (Figure
was 24.6°C), which indicated maintenance of aerobic 2). The pH increase in the inoculated silages during the
stability. This indicates that exposure of the silage aerobic stability test was not related to the higher yeast
to air for a period of 12 h did not promote changes growth because the inoculated silages presented the
that compromised the aerobic stability of the material, lowest yeast counts at the end of the test. Inoculated
which was also evidenced by the similar microorganism silages had a higher pH from the opening of the silo
count among all treatments at silo opening (Table 3). and consequently a higher pH at the end of the stabil-
The pH of the silage during the re-ensiling process is ity test. However, when we observe Figure 2, the pH
an important factor for the success of the operation pattern is very similar between inoculated and nonin-
because some microorganisms deteriorating the silage, oculated silages. The expected behavior would be an
such as molds, do not develop in an acidic environment increase in yeast counts under low-pH (3.5–4.0) aerobic

Journal of Dairy Science Vol. 101 No. 7, 2018

 RE-ENSILING SORGHUM SILAGE 6053

CONCLUSIONS

Re-ensiling after 12 h of aerobic exposure followed

by 240 d of anaerobic storage did not lead to reduced
nutritional quality or loss of aerobic stability compared
with silages that remained sealed throughout storage.
In our study, the inoculant did not contribute measur-
ably to the success of the re-ensiling process.

ACKNOWLEDGMENTS

The authors thank the Brazilian Agricultural Research

Corporation (EMBRAPA)—Maize and Sorghum, Sete
Lagoas, Minas Gerais, Brazil, the Coordination and
Improvement of Higher Level or Education Personnel
Figure 1. Changes in pH and SEM during air exposure during (CAPES), the Veterinary School of Federal University
evaluation of aerobic stability. Average of all treatments.
of Minas Gerais (EVUFMG), and the Pro-Rector of
Research of Federal University of Minas Gerais (PRPq-
UFMG), Belo Horizonte, Minas Gerais, Brazil, for their
support and assistance.
conditions. In our experiment, the pH of the inoculated
silages was 4.4, so other aerobic microorganisms may
have been favored. Thus, the lower yeast count can be REFERENCES
explained by lower growth rate in relation to aerobic AOAC International. 2005. Official Methods of Analysis. 18th ed.
bacteria (Muck, 2010). Although the aerobic bacteria AOAC International, Gaithersburg, MD.
count did not show statistical differences, counts were Chen, Y., and Z. G. Weinberg. 2014. The effect of relocation of whole-
crop wheat and corn silages on their quality. J. Dairy Sci. 97:406–
high in all treatments. 410.
Overall, our results agree with the few studies on de Lima, E. M., L. C. Gonçalves, K. M. Keller, J. A. dos Santos Ro-
re-ensiling published so far (Chen and Weinberg, 2014; drigues, F. P. C. Santos, P. H. F. Michel, V. S. Raposo, and D.
G. Jayme. 2016. Re-ensiling and its effects on chemical composi-
de Lima et al., 2016; Michel et al., 2017) that the re-en- tion, in vitro digestibility, and quality of corn silage after different
siling process does not significantly affect silage quality. lengths of exposure to air. Can. J. Anim. Sci. 97:250–257.
However, it is important that aspects such as the qual- Filya, I., G. Ashbell, Y. Hen, and Z. G. Weinberg. 2000. The effect of
bacterial inoculants on the fermentation and aerobic stability of
ity of the silage to be relocated are taken into account whole crop wheat silage. Anim. Feed Sci. Technol. 88:39–46.
so that the risks of loss in the process are minimized. Filya, I., E. Sucu, and A. Karabulut. 2004. The effect of Propionibac-
terium acidipropionici, with or without Lactobacillus plantarum,
on the fermentation and aerobic stability of wheat, sorghum and
maize silages. J. Appl. Microbiol. 97:818–826.
Filya, I., E. Sucu, and A. Karabulut. 2006. The effects of Propioni-
bacterium acidipropionici and Lactobacillus plantarum, applied
at ensiling, on the fermentation and aerobic stability of low dry
matter corn and sorghum silages. J. Ind. Microbiol. Biotechnol.
33:353–358.
Holden, L. A. 1999. Comparison of methods of in vitro dry matter
digestibility for ten feeds. J. Dairy Sci. 82:1791–1794.
Jobim, C. C., L. G. Nussio, R. A. Reis, and P. Schmidt. 2007. Meth-
odological advances in evaluation of preserved forage quality. Rev.
Bras. Zootec. 36:101–119.
Kristensen, N. B., K. H. Sloth, O. Højberg, N. H. Spliid, C. Jensesn,
and R. Thøgersen. 2010. Effects of microbial inoculants on corn
silage fermentation, microbial contents, aerobic stability, and milk
production under field conditions. J. Dairy Sci. 93:3764–3774.
McDonald, P. A. R. H., and S. J. E. Heron. 1991. The Biochemistry of
Silage. 2nd ed. Chalcombe, Bucks, UK.
Michel, P. H. F., L. C. Gonçalves, J. A. S. Rodrigues, K. M. Keller, V.
S. Raposo, E. M. Lima, F. P. C. Santos, and D. G. Jayme. 2017.
Re-ensiling and inoculant application with Lactobacillus plantarum
and Propionobacterium acidipropionic on sorghum silages. Grass
Forage Sci. 72:432–440.
Figure 2. Changes in pH and SEM of silages with and without Muck, R. E. 2010. Silage microbiology and its control through addi-
inoculant. tives. Rev. Bras. Zootec. 39:183–191.

Journal of Dairy Science Vol. 101 No. 7, 2018

6054 DOS ANJOS ET AL.

National Research Council. 2001. Nutrient Requirements of Dairy Silva, D. J., and A. C. Queiroz. 2002. Análise de Alimentos (Métodos
Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC. Químicos e Biológicos). 3rd ed. UFV, Viçosa, MG, Brazil.
Pahlow, G., R. E. Muck, F. Driehuis, S. J. W. H. Oude Elferink, and Tabacco, E., S. Piano, L. Cavallarin, T. F. Bernardes, and G. Bor-
S. F. Spoelstra. 2003. Microbiology. Pages 31–93 in Silage Science reani. 2009. Clostridia spore formation during aerobic deteriora-
and Technology, Agronomy Monograph 42. D. R. Buxton, R. E. tion of maize and sorghum silages as influenced by Lactobacillus
Muck, and J. H. Harrison, ed. American Society of Agronomy, buchneri and Lactobacillus plantarum inoculants. J. Appl. Micro-
Madison, WI. biol. 107:1632–1641.
Pitt, J. I., and A. D. Hocking. 2009. Fungi and Food Spoilage. 3rd ed. Tabacco, E., F. Righi, A. Quarantelli, and G. Borreani. 2011. Dry
Springer, New York, NY. matter and nutritional losses during aerobic deterioration of corn
Pitt, R. E., and R. E. Muck. 1993. A diffusion model of aerobic dete- and sorghum silages as influenced by different lactic acid bacteria
rioration at the exposed face of bunker silos. J. Agric. Eng. Res. inocula. J. Dairy Sci. 94:1409–1419.
55:11–26. Tilley, J. M. A., and R. A. Terry. 1963. A two-stage technique for the
Playne, M. J. 1985. Determination of ethanol, volatile fatty acids, in vitro digestion of forage crops. Grass Forage Sci. 18:104–111.
lactic and succinic acids in fermentation liquids by gas chromatog- Van Soest, P. J., J. B. Robertson, and B. A. Lewis. 1991. Methods for
raphy. J. Sci. Food Agric. 36:638–644. dietary fiber, neutral detergent fiber, and non starch polysaccha-
Ranjit, N. K., and L. Kung Jr. 2000. The effect of Lactobacillus bu- rides in relation to animal nutrition. J. Dairy Sci. 74:3583–3597.
chneri, Lactobacillus plantarum, or a chemical preservative on the
fermentation and aerobic stability of corn silage. J. Dairy Sci.
83:526–535.

Journal of Dairy Science Vol. 101 No. 7, 2018