Full text of DOI:10.3168/jds.S0022-0302(01)74666-8

J. Dairy Sci. 84:2195–2202
 American Dairy Science Association, 2001.
The Effect of a Dry or Liquid Application of Lactobacillus
plantarum MTD1 on the Fermentation of Alfalfa Silage1
A. G. Whiter and L. Kung, Jr.
Delaware Agricultural Experiment Station
Department of Animal and Food Science
College of Agriculture and Natural Resources
University of Delaware
Newark, 19717-1303
INTRODUCTION
ABSTRACT
Homolactic bacteria have been used to inoculate si-
lages and improve their fermentation (Kung et al.,
1984; Stokes, 1992). Inoculants are usually supplied as
freeze-dried powders for use either suspended in water
for spray application, or as granules for dry application.
For application in water, the freeze-dried bacteria re-
suscitate primarily in the water used for application.
In contrast, inoculants applied in a dry form rely solely
on moisture in or on the crop to resuscitate the organ-
isms. Thus, it may take longer for the bacteria in an
inoculant applied in a dry form to revive, giving a slower
initial rate of fermentation than with an inoculant ap-
plied in water. The microorganisms used in silage inoc-
ulants depend on a high level of moisture for growth
(Lanigan, 1963). Water activity (aw), rather than DM
or moisture content, is a more accurate way to deter-
mine how much water in a feedstuff is available for
microbial growth during fermentation (Albert et al.,
1989). Water activity has been shown to affect all four
bacterial growth phases: length of lag phase, growth
rate, size of stationary population, and death rate
(Toller and Christian, 1978). Pahlow and Weissbach
(1996) reported that only 10% of the total population
of epiphytic lactic acid bacteria (LAB) on grasses grew
on modified, selective agar with an aw of 0.952 (corres-
ponding to forage with a DM content of 50%), suggesting
that epiphytic LAB were not very osmotolerant. How-
Alfalfa was wilted to 30 and 54% dry matter and
was untreated or treated with Lactobacillus plantarum
MTD1 applied either as a liquid or dry inoculant to
determine whether form of inoculation affected silage
fermentation. In silages with 30% dry matter, both
forms of inoculation resulted in silages with more lactic
acid and a lower pH than in untreated silage after 2 d
of ensiling. In addition, both forms of inoculation re-
sulted in silages with lower concentrations of acetic acid
between 8 and 45 d of ensiling than in untreated silage.
The ammonia-N content was also lower in silages that
had been treated with both forms of inoculation during
the early and intermediate stages of ensiling, but not
after 45 d of ensiling. In silages containing 54% dry
matter, dry and liquid inoculation produced a more
rapid decrease in pH from d 4 to 14 when compared
with untreated silage, but the effect was greater when
the inoculant was applied in water. In contrast to find-
ings from 30% silages, the inoculated silages with 54%
dry matter had lower concentrations of ammonia-N
than untreated silage from d 8 to 45, and both forms
of inoculation were equally effective. The results of this
study agree with previous research that shows that
microbial inoculation can improve the fermentation of
alfalfa silage. However, this report shows that a micro-
bial inoculant was more effective when applied in a
liquid- rather than a dry-form to alfalfa with a high ever, alfalfa is often wilted to concentrations above 30%
DM before ensiling to reduce the chances of a clostridial
fermentation (Leibensperger and Pitt, 1987). Further-
more, conditions in some geographical areas for wilting
silage are particularly good, and it is common to find
silages containing more than 45% DM. In these silages,
it may be speculated that the reduced aw may further
challenge revival of an inoculant applied in a dry form
unless the bacteria are osmotolerant (able to withstand
high osmotic pressures) and are able to grow quickly
when the aw is low.
DM content.
(Key words: inoculant, silage, fermentation)
Abbreviation key: aw = water activity, LAB = lactic
acid bacteria, WSC = water-soluble carbohydrates.
Received March 19, 2001.
Accepted May 16, 2001.
Corresponding author: L. Kung, Jr.; e-mail: lkung@udel.edu.
1Published as paper number 1697 in the Journal series of the
Delaware Agricultural Experiment Station.
The objective of this study was to determine the effi-
cacy of a single strain of L. plantarum applied in a
2195

2196
WHITER AND KUNG, JR.
liquid or dry form on the fermentation of alfalfa ensiled
The numbers of LAB were determined by pour plating
at a moderate (about 30%) and high (about 54%) DM on de Mann Rogosa Sharpe agar. Sterile cyclohexamide
content.
was added after autoclaving to obtain a concentration
of 0.01% wt/vol. Plates were incubated anaerobically
at 32°C for 48 h. The numbers of enterobacteria were
enumerated by pour plating on violet red bile glucose
agar (Oxoid CM485) with a single overlay. Plates were
incubated under aerobic conditions at 36°C for 18 h.
Yeast and molds were determined by pour plating in
malt extract agar (Oxoid CM59) that had been acidified,
after autoclaving, by the addition of 85% lactic acid at
a concentration of 0.5% vol/vol. Plates were incubated
under the same conditions as those for LAB.
For analysis of acetic, propionic, and butyric acids, a
portion of the water extract was diluted 5:1 with 25%
meta-phosphoric acid containing isocaproic acid as an
internal standard. The resulting mixture was centri-
fuged at 10,000 × g for 10 min before injecting into
a Hewlett Packard 5890 gas chromatograph (Hewlett
Packard, Avondale, PA) with a 530-µm (internal diame-
ter) Carbowax 20-m column (Supelco, Bellefonte, PA).
The chromatograph oven was programmed as follows:
70°C for 1 min, 5°C increase/min to 100°C, 45°C in-
crease/min to 170°C, and a final holding time of 5 min.
Ammonia-N was analyzed on the water extracts by a
phenol-hypochlorite procedure as described by Weath-
erburn (1967). Water-soluble carbohydrates (WSC)
were analyzed by the procedure described by Nelson
MATERIALS AND METHODS
Silo Study
Alfalfa was mowed during second cutting at the early
bloom stage of maturity. After wilting to 30% DM, every
other row of forage was chopped by a conventional for-
age harvester to a theoretical length of 0.95 cm. The
remaining forage continued to wilt to obtain a DM con-
centration of 54% before chopping. Two formulations,
one for a dry granular application and the other for a
liquid application, of a microbial inoculant containing a
single strain of Lactobacillus plantarum-MTD1 (Ecosyl
Products, Ltd., Stokesley, England) were enumerated
on de Mann Rogosa Sharpe agar (Oxoid CM361, Oxoid,
Basingstoke, UK). Based on the concentration of LAB
in the inoculants, appropriate amounts were weighed
out to supply 1 × 105 cfu/g of wet forage. Chopped forages
(30 and 54% DM) were divided into three piles each
and treated with 1) 2 L of deionized water/tonne, un-
treated, 2) the dry granular form of inoculant followed
by the addition of 2 L of deionized water/tonne, and 3)
the inoculant applied in deionized water at 2 L/tonne.
The dry inoculant formulation was sprinkled by hand
onto the forage mass while the sample was mixed. Wa-
ter alone or containing the inoculant was applied with
a hand-sprayer while mixing the forage samples. After
treatment, forages were packed into laboratory silos
(0.203 × 0.064 m) and ensiled for 2, 4, 8, 14, or 45 d
between 22 to 25°C. Three silos for each treatment were
opened at each sampling time. Packing density was
approximately 250 kg of DM/m3. Packing and sealing
of silos was completed within 1 to 2 h of chopping.
Forages were harvested and chopped on sunny days
without rain.
(1944).
cedure (kit 826-UV, Sigma-Aldrich, St. Louis, MO). For
the analysis of -lactic acid, -lactic dehydrogenase was
replaced with a similar amount of -lactic dehydroge-
nase (Sigma -9636). -Lactic acid (Sigma L-2250) and
-lactic acid (Sigma L-1000) were used as standards for
their respective assays. Lactic acid is reported as the
sum of - and -lactic acid.
L
-Lactic acid was analyzed by an enzymatic pro-
D
L
D
L
L
D
L
D
Osmotolerance of the Microbial Inoculant
Before ensiling, samples were obtained after the addi-
tion of each treatment and stored on ice until they were
returned to the lab for processing. The DM content of
fresh forages and silages was determined by drying at
60°C for 48 h in a forced-draft oven. Samples of forages
and silages (25 g) were homogenized in 225 ml of sterile
quarter-strength Ringer’s solution (Oxoid BR52) for 1
min and then, filtered through Whatman 54 filter paper
(Whatman, Inc., Clifton, NJ). The pH of the homoge-
nized solution was determined with a Corning pH meter
(model no. 220, Corning Science Products, Corning,
NY). The filtered water extract was acidified with 15
µl of 50% (wt/vol) H2SO4 to reduce the pH of the extract
to < 2.0 before freezing. The homogenized samples were
serially diluted (10-fold) for the enumeration of LAB,
enterobacteria, yeasts, and molds.
Potassium chloride was added to de Mann Rogosa
Sharpe agar. The varying amounts of KCl in the agar
produced five treatments with aw that corresponded to
forages with a range of DM between 0 and 50% as
described by Weissbach (1968). Sterile deionized water
was added to the molten agar and KCl mixture to com-
pensate for water lost during autoclaving. Five grams
of the dry-granular inoculant formulation was diluted
in 45-ml sterile quarter-strength Ringer’s solution (Ox-
oid BR52) and allowed to stand for 15 min. This mixture
was then serially diluted (10-fold) and pour-plated in
the prepared agars in duplicate for each of the five KCl
treatments. The plates were sealed with Parafilm to
prevent loss of moisture and incubated at 37°C for 48
h. To determine the aw of the agars, 3.5 ml of each agar
Journal of Dairy Science Vol. 84, No. 10, 2001

SILAGE FERMENTATION
2197
Table 1. The DM content (%) of alfalfa forage and silage treated with different forms of a microbial inoculant.
Days of ensiling
Treatment
0
2
4
8
14
45
30% DM
Control
29.0b
29.7b
30.9b
29.8c
30.4c
30.9c
29.4d
29.9d
31.4c
30.1d
30.4d
31.8c
29.9d
30.4d
31.8c
30.1b
30.6b
31.3b
Liquid inoculant1
Dry inoculant2
54%DM
Control
Liquid inoculant
Dry inoculant
54.0a
53.6a
54.1a
53.3b
54.9a
54.5a
52.6b
54.5a
54.5a
52.9b
54.1ab
54.9a
53.0b
54.1a
54.4a
52.4a
53.4a
53.5a
SE
0.4
0.2
0.3
0.2
0.2
0.4
a,bMeans in columns with unlike superscript differ (P < 0.05).
1Lactobacillus plantarum MTD1 (Ecosyl Products, Ltd., Stokesley, England) applied in a liquid form to
achieve a final concentration of 1 × 105 cfu/g of fresh forage.
2Lactobacillus plantarum MTD1 applied in a dry form to achieve a final concentration of 1 × 105 cfu/g of
fresh forage.
was placed in a sample cup and allowed to stand for from 5.04 to 5.41 log cfu/g of forage (Table 2). At 2 d of
30 min. The aw of the agars were then measured with
a water activity meter (model CX-1, Decagon, Pullman,
WA). Four measurements were taken for each of the
five agar treatments.
ensiling, the number of LAB in silages with 30% DM
had markedly increased to a range of 8.86 to 9.49 log
cfu/g. However, inoculated silages did not have greater
numbers of LAB than untreated silage after 4 d of ensil-
ing in 30% DM silages. In silages containing 54% DM,
numbers of LAB increased to a lesser (P < 0.05) extent
after 2 d of ensiling when compared with the increase
Statistical Analysis
All microbial data were transformed to log10. Chemi- in numbers of LAB in 30% DM forage. Inoculation as
cal data are presented on a DM basis. Data from the
laboratory silo study were analyzed using the general
linear models procedure of SAS (1998) for a completely
randomized design. The statistical model used in the
analysis of the fermentation data was:
a liquid, but not dry treatment (P < 0.05), increased the
numbers of LAB compared with untreated silage at this
time. Silage treated with the inoculant applied as a
liquid had higher numbers of LAB than untreated si-
lage or silage treated with a dry-granular inoculant
through 8 d of fermentation.
Changes in numbers of enterobacteria during ensiling
are shown in Table 3. In silage containing 30% DM,
application of the liquid inoculant resulted in a more
rapid decline in enterobacteria than in untreated silages
(P < 0.05) after 2 d of ensiling. Similarly, silage with 54%
DM and treated with the liquid form of inoculant had
lower numbers of enterobacteria than untreated silage,
but this did not occur until the d 8 of ensiling. In contrast,
inoculation with the dry-granular form of inoculant, at
both DM, did not affect the numbers of enterobacteria
during ensiling when compared to untreated silages.
Yeasts (data not shown, <2.00 log10cfu/g) and molds (data
not shown, <3.00 log10cfu/g) were unaffected by inocula-
tion or DM content of the silage.
Yijkl = µ + Di + Ij + DIij + Tk + DTik + ITjk + DITijk
+ εijkl
,
where Yijkl is an observation, µ is the population mean,
Di is the effect of DM i, Ij is the effect of inoculation j,
DIij is the interaction between DM and inoculation ij,
Tk is the effect of time k, DTik is the interaction between
DM and time ik, ITjk is the interaction between inocula-
tion and time jk, DITijk is the interaction among DM,
inoculation and time ijk, and εijkl is the residual error.
Differences among means were tested using Tukey’s
Test (Snedecor and Cochran, 1980). An α level of P <
0.05 was deemed significant.
The concentrations of WSC decreased more rapidly
(P < 0.05) between 4 to 45 d of ensiling in 30% than in
54% DM silages (Table 4). Inoculated silages (dry and
liquid) containing 30% DM had lower (P < 0.05) concen-
trations of WSC after 2 d of ensiling than did untreated
RESULTS
The DM content of forages was consistent throughout
the ensiling period, with the lower DM and higher DM
silages averaging 30 and 54%, respectively (Table 1).
Fresh forages contained numbers of LAB that ranged silage with 30% DM. However, thereafter, WSC concen-
Journal of Dairy Science Vol. 84, No. 10, 2001

2198
WHITER AND KUNG, JR.
Table 2. Numbers of lactic acid bacteria (log cfu/g) in alfalfa forage and on silages treated with different
forms of a microbial inoculant.
Days of ensiling
Treatment
0
2
4
8
14
30% DM
Control
5.04
5.29
5.56
8.86b
9.49a
9.11ab
9.66a
9.68a
9.57a
9.08b
9.13b
8.93b
8.93bc
8.98b
8.69d
Liquid inoculant1
Dry inoculant2
54% DM
Control
Liquid inculant
Dry inoculant
5.41
5.16
5.39
7.55d
8.09c
7.73cd
8.74c
9.54a
9.08b
8.93b
9.56a
9.17b
9.08b
8.77cd
9.28a
SE
0.31
0.08
0.06
0.08
0.04
a,b,c,dMeans in columns with unlike superscript differ (P < 0.05).
1Lactobacillus plantarum MTD1 (Ecosyl Products, Ltd., Stokesley, England) applied in a liquid form to
achieve a final concentration of 1 × 105 cfu/g of fresh forage.
2Lactobacillus plantarum MTD1 applied in a dry form to achieve a final concentration of 1 × 105 cfu/g of
fresh forage.
Table 3. The number of enterobacteria (log10 cfu/g) on alfalfa forage and silage treated with different forms
of a microbial inoculant.
Days of ensiling
Treatment
0
2
4
8
14
30% DM
Control
5.60b
7.10a
5.53b
6.03ab
4.31cd
3.14d
3.37cd
3.34b
3.49b
3.24b
3.33
3.18
2.89
Liquid inoculant1
Dry inoculant2
54% DM
Control
Liquid inoculant
Dry inoculant
6.07ab
6.16ab
6.32ab
6.79a
6.68a
5.78ab
4.57bc
5.94a
4.64a
3.11b
4.61a
2.93
3.10
1.97
6.17ab
6.35ab
6.21ab
SE
0.20
0.24
0.26
0.21
0.45
a,b,c,dMeans in columns with unlike superscript differ (P < 0.05).
1Lactobacillus plantarum MTD1 (Ecosyl Products, Ltd., Stokesley, England) applied in a liquid form to
achieve a final concentration of 1 × 105 cfu/g of fresh forage.
2Lactobacillus plantarum MTD1 applied in a dry form to achieve a final concentration of 1 × 105 cfu/g of
fresh forage.
Table 4. The concentrations of water soluble carbohydrates (% DM) in alfalfa silages treated with different
forms of a microbial inoculant.
Days of ensiling
Treatment
0
2
4
8
14
45
30% DM
Control
8.67a
8.57a
8.73a
8.87a
3.37b
3.80b
2.43c
1.87d
2.63d
2.50d
0.87d
0.90d
2.13d
1.13c
0.87c
0.90c
Liquid inoculant1
Dry inoculant2
54% DM
Control
Liquid inoculant
Dry inoculant
2.57c
2.57c
5.50c
8.60a
10.83a
7.30b
9.93a
9.83a
4.60c
7.60b
7.70a
4.23c
5.87b
5.57a
4.20b
3.90b
7.40ab
6.90bc
10.80a
10.87a
SE
0.31
0.94
0.31
0.25
0.34
0.25
a,b,c,dMeans in columns with unlike superscript differ (P < 0.05).
1Lactobacillus plantarum MTD1 (Ecosyl Products, Ltd., Stokesley, England) applied in a liquid form to
achieve a final concentration of 1 × 105 cfu/g of fresh forage.
2Lactobacillus plantarum MTD1 applied in a dry form to achieve a final concentration of 1 × 105 cfu/g of
fresh forage.
Journal of Dairy Science Vol. 84, No. 10, 2001

SILAGE FERMENTATION
2199
After 2 d of ensiling, and in treated silages with a
DM content of 30%, both forms of inoculation had more
(P < 0.05) than double the concentration of lactic acid
(5.66 to 5.79%) than in untreated silage (2.50%, Table
5). However, the lactic acid concentration was similar
among treatments at d 4 to 45. For the 54% DM silages,
the concentration of lactic acid in silage treated with
the dry form of inoculant was not different from the
lactic acid concentration in untreated silage at any time
during the fermentation although it was numerically
greater in inoculated silages. In contrast, the liquid
method of inoculation resulted in silage with a higher
concentration (P < 0.05) of lactic acid than in untreated
silage after 8 and 14 d of ensiling. After 45 d of ensiling,
there were no differences among treatments in lactic
acid content within either DM content, but the concen-
trations of lactic acid were higher in the silages with
30% than 54% DM content.
The acetic acid concentration was much lower (P <
0.05) in the 54% DM silages (0.40 to 0.59%) than in the
30% DM silages (2.54 to 3.10%) with no effects due to
inoculation in the former after 45 d of ensiling (Table
6). In silages with 30% DM content, both forms of inocu-
lation decreased (P < 0.05) the production of acetic acid
at 8, 14, and 45 d of ensiling when compared with un-
treated silage. The dry form of inoculation resulted in
lower (P < 0.05) concentrations of acetic acid than in
silages treated with the liquid form of inoculant after
8 and 14 d of ensiling, but this difference was not detect-
able after 45 d of ensiling. Treatment did not affect the
concentration of acetic acid in 54% DM silage. Propionic
and butyric acids were less than 0.10% of the DM in all
of the silages and were not different among treatments
(data not shown).
At both DM concentrations, the addition of liquid
and dry inoculants decreased (P < 0.05) the amount of
ammonia-N relative to untreated silages at 8 and 14 d
of ensiling (Table 7). In silages with 30% DM, the dry
inoculant was more effective (P < 0.05) than the liquid
inoculant. At the end of ensiling (45 d), concentrations
of ammonia-N in silages with 54% DM (0.66 to 0.126%)
were more (P < 0.05) than half that of the ammonia-
N in the silages with the lower DM content (0.242 to
0.292%). Inoculation reduced the ammonia-N concen-
tration only in the higher DM silages (P < 0.05) after
45 d of ensiling.
Figure 1.The effect of form of application of a microbial inoculant
on the decrease in pH of low DM (A) and high DM (B) alfalfa silage.
Untreated silage (᭹), silage Lactobacillus plantarum MTD1 (Ecosyl
Products, Ltd., Stokesley, England) in a dry form (᭝), silage treated
Lactobacillus plantarum MTD1 in a liquid form (᭿). SE < 0.05.
trations were not different among these treatments. In
contrast, in the silages with 54% DM, inoculation (dry
and liquid) resulted in silage with lower (P < 0.05) con-
centrations of WSC than in untreated silage from d 8
to 45. The concentrations of WSC decreased more rap-
idly (P < 0.05) in silage treated with the liquid inoculant
than the dry inoculant between d 4 and 14.
Fresh forages had a pH of about 6.0 (Figures 1A and
1B). After 2 d of ensiling, silages with 30% DM and
treated with both forms of inoculation had lower (P <
0.05) pH than untreated silages, but after 4 d of ensil-
ing, all silages had a similar pH (about 4.4). However,
in the higher DM forage, differences in pH were not
observed between untreated and treated silages until
the fourth day of ensiling. Treating with both forms of
the inoculant resulted in lower (P < 0.05) silage pH,
but the liquid form of inoculant was more effective (P
< 0.05) than the dry form of inoculant between d 4 and
8. The final pH of silage with 54% DM was lower (P <
0.05) in inoculated silage than in untreated silage after
45 d of ensiling.
The amounts of KCl added to obtain Rogosa agar
with different theoretical aw are shown in Table 8. The
actual measured aw were slightly lower than theoretical
aw values. The numbers of culturable LAB were reduced
as the aw decreased from 0.987 to 0.949. No growth was
detected on agar with an aw of 0.942.
Journal of Dairy Science Vol. 84, No. 10, 2001

2200
WHITER AND KUNG, JR.
Table 5. The lactic acid concentration (% DM) of alfalfa silages treated with different forms of a microbial
inoculant.
Days of ensiling
Treatment
2
4
8
14
45
30% DM
Control
2.50b
5.79a
5.66a
6.13a
7.98a
6.18a
6.35ab
6.79a
7.18a
7.47a
7.00a
8.21a
7.89a
7.71a
Liquid inoculant1
Dry inoculant2
54% DM
Control
Liquid inoculant
Dry inoculant
6.41ab
1.08c
1.04c
0.92c
1.21b
3.23b
1.64b
2.46c
5.47b
3.29c
4.32c
5.92b
5.00bc
5.09b
4.90b
4.75b
SE
0.12
0.48
0.27
0.23
0.22
a,b,cMeans in columns with unlike superscript differ (P < 0.05).
1Lactobacillus plantarum MTD1 (Ecosyl Products, Ltd., Stokesley, England) applied in a liquid form to
achieve a final concentration of 1 × 105 cfu/g of fresh forage.
2Lactobacillus plantarum MTD1 applied in a dry form to achieve a final concentration of 1 × 105 cfu/g of
fresh forage.
Table 6. The acetic acid concentration (% DM) of alfalfa silages treated with different forms of a microbial
inoculant.
Days of ensiling
Treatment
2
4
8
14
45
30% DM
Control
1.70a
1.41a
1.48a
1.70a
1.41a
1.48a
2.39a
1.83b
1.45c
2.24a
1.84b
1.38c
3.10a
2.67b
2.54b
Liquid inoculant1
Dry inoculant2
54% DM
Control
Liquid inoculant
Dry inoculant
0.36b
0.34b
0.29b
0.36b
0.34b
0.29b
0.39d
0.36d
0.29d
0.40d
0.32d
0.34d
0.59c
0.40c
0.42c
SE
0.01
0.01
0.007
0.003
0.01
a,b,c,dMeans in columns with unlike superscript differ (P < 0.05).
1Lactobacillus plantarum MTD1 (Ecosyl Products, Ltd., Stokesley, England) applied in a liquid form to
achieve a final concentration of 1 × 105 cfu/g of fresh forage.
2Lactobacillus plantarum MTD1 applied in a dry form to achieve a final concentration of 1 × 105 cfu/g of
fresh forage.
Table 7. The ammonia-N concentrations (% DM) of alfalfa silages treated with different forms of a microbial
inoculant.
Days of ensiling
Treatment
0
2
4
8
14
45
30% DM
Control
0.022ab
0.021ab
0.029a
0.087a
0.062b
0.062b
0.134a
0.095b
0.075c
0.172a
0.132b
0.105c
0.238a
0.208b
0.147c
0.292a
0.263ab
0.242b
Liquid inoculant1
Dry inoculant2
54% DM
Control
Liquid inoculant
Dry inoculant
0.018b
0.018b
0.021ab
0.052bc
0.036d
0.041cd
0.063cd
0.043e
0.050de
0.085d
0.048e
0.059e
0.108d
0.057e
0.070e
0.126c
0.066d
0.084d
SE
0.002
0.002
0.004
0.003
0.004
0.006
a,b,c,d,eMeans in columns with unlike superscript differ (P < 0.05).
1Lactobacillus plantarum MTD1 (Ecosyl Products, Ltd., Stokesley, England) applied in a liquid form to
achieve a final concentration of 1 × 105 cfu/g of fresh forage.
2Lactobacillus plantarum MTD1 applied in a dry form to achieve a final concentration of 1 × 105 cfu/g of
fresh forage.
Journal of Dairy Science Vol. 84, No. 10, 2001

SILAGE FERMENTATION
2201
Table 8. The effect of adding potassium chloride to de Mann Rogosa Sharpe agar to vary the water activity
(aw) of the agar on numbers of culturable Lactobacillus plantarum.
Potassium
chloride, g/L of
de Mann Rogosa
Sharpe agar1
Culturable lactic
acid bacteria,
log10 cfu/g
1
2
Theoretical aw
Measured aw
0
60
83.2
97.4
113
0.990
0.970
0.960
0.956
0.952
0.987
0.964
0.953
0.949
0.942
9.5
9.45
9.22
9.18
0
1From Weissbach, 1968.
2Using a water activity meter.
DISCUSSION
and Weissbach (1999) reported that the addition of an
inoculant in water was more effective than a dry appli-
cation of the same inoculant in lowering the pH of wilted
(45% DM) grass silage. In the future, selection of micro-
organisms for use in high DM silages should include
some measurement of osmotolerance. The current Ger-
man scheme for approval of silage additives includes
such a measurement (Pahlow and Weissbach, 1999).
Both forms of inoculation in the lower, but not higher,
DM silages markedly decreased the production of acetic
acid during the later stages of fermentation. Inoculation
with homolactic LAB also did not effect the concentra-
tion of acetic acid in alfalfa silages with high concentra-
tions of DM (Keller et al., 1994; Kung et al. 1987). The
concentration of acetic acid in the higher DM silage
was markedly lower than in the 30% DM silage after
45 d of ensiling in our study. Thus, the ratio of lactic
to acetic acid ranged from 2.6 to 3.0 in the lower DM
silage, but the ratio was greater than 8.5 in the higher
DM silages. In addition, the concentration of acetic acid
did not change appreciably between 2 and 45 d in silages
with the higher concentration of DM. This suggests that
microorganisms responsible for acetic acid production
in these silages were limited to growth in the very early
stages of ensiling. Lack of differences among silage
treatments with 54% DM may be related to these obser-
vations.
Silage fermentation proceeds at a slower rate and
usually to a lesser extent with increasing DM content
(Jones et al., 1992; Muck, 1990) because the low aw
restricts the growth of bacteria (Toller and Christian,
1978). Results from the current study showed that num-
bers of LAB did not increase as rapidly and lactic acid
concentrations were lower in silages with the higher
DM content. Consequently, the pH did not decrease as
rapidly in silage with 54% DM as in silage with 30%
DM content. We also found that growth of L. plantarum
MTD1, used in our inoculant, was reduced when the
aw of modified media was decreased from 0.987 to 0.949.
In a model of lactic acid fermentation, Pitt et al. (1985)
reported that the relative growth of LAB would have
been decreased by more than 50% between these aw.
Furthermore, we observed that colony sizes (data not
reported) were also smaller (an indication of slower
growth rate) when the aw was ≤ 0.953. Pahlow and
Weissbach (1996) also reported smaller colony sizes in
LAB grown on media with a low aw content. They also
observed that only 10% of epiphytic LAB from grass
were able to grow on selective media with aw of 0.952,
which is equivalent to forage of about 50% DM. In our
study, the strain of L. plantarum was highly osmotoler-
ant because colonies were detectable even when the aw
of the media was 0.949.
Microbial inoculation with homofermentative LAB is
also beneficial because a more rapid fall in silage pH
can result in less proteolysis and deamination, and
therefore lower concentrations of ammonia-N. Micro-
bial inoculation reduced the accumulation of ammonia-
N in the higher DM silage between d 8 to 45. Using the
same species of L. plantarum as in the current study,
Kung et al. (1991) reported similar findings in alfalfa
with a 45% DM content. The degree in the reduction
of ammonia-N was not affected by form of inoculation,
although one might have expected the liquid applica-
tion of the inoculant to be better because it had a more
Inoculation as a dry or liquid product on alfalfa with
30% DM, equally stimulated the early growth of LAB
and caused a more rapid decline in silage pH after 2 d
of ensiling. Similar effects were found in the production
of lactic acid, a direct consequence of the changes in
number of LAB. In contrast, in silage with the higher
DM content, application of the inoculant in water re-
sulted in greater numbers of LAB than did application
in a dry form. The inoculant applied as a liquid also
resulted in silage with less enterobacteria, a more ex-
tensive decline in pH, and a greater production of lactic
acid compared with the inoculant applied in a dry form
after 4 d of ensiling. Similar to our findings, Pahlow marked effect on lowering silage pH. However, silage
Journal of Dairy Science Vol. 84, No. 10, 2001

2202
WHITER AND KUNG, JR.
treated with the liquid inoculant did have numerically, cal assistance from N. K. Ranjit, J. A. Mills, J. R. Rob-
but not statistically, lower concentrations of ammonia- inson, and J. Y. Tavares. A. G. Whiter was a Science &
N than silage that had been treated with the dry in- Engineering Scholar who completed this research as a
oculant.
partial fulfillment for a Degree with Distinction.
REFERENCES
CONCLUSIONS
Albert, R. A., B. Huebner, and L. W. Davis. 1989. Role of water
activity in the spoilage of alfalfa hay. J. Dairy Sci. 72:2573–2581.
Jones, B. A., L. D. Satter, and R. E. Muck. 1992. Influence of bacterial
inoculant and substrate addition to lucerne ensiled at different
dry matter contents. Grass Forage Sci. 47:19–27.
Keller, T., H. Nonn, and H. Jeroch. 1994. Comparative studies on
the efficiency of various biological silage additives for the ensiling
of lucerne. Arch. Anim. Nutr. 47:75–87.
Kung, L., Jr., D. B. Grieve, J. W. Thomas, and J. T. Huber. 1984.
Added ammonia or microbial inocula for fermentation and nitrog-
enous compounds of alfalfa ensiled at various percents of dry
matter. J. Dairy Sci. 67:299–306.
Kung, L., Jr., R. S. Tung, and K. Maciorowski. 1991. Effect of a
microbial inoculant and/or glycopeptide antibiotic on fermenta-
tion and aerobic stability of wilted alfalfa silage. Anim. Feed Sci.
Tech. 35:37–48.
Kung, L., Jr., L. D. Satter, B. A. Jones, K. W. Genin, A. L. Sudoma,
G. L. Enders, and H. S. Kim. 1987. Microbial inoculation of low
moisture alfalfa silage. J. Dairy Sci. 70:2069–2077.
Lanigan, G. W. 1963. Silage bacteriology. I. Water activity and tem-
perature relationships of silage strains of Lactobacillus plan-
tarum, Lactobacillus brevis, and Pediococcus cerevisiae. Aust. J.
Biol. Sci. 16:606–615.
Leibensperger, R. Y., and R. E. Pitt. 1987. A model of clostridial
dominance in ensilage. Grass Forage Sci. 42:297–317.
Muck, R. E. 1990. Dry matter level effects on alfalfa silage quality
II. Fermentation products and starch hydrolysis. Trans. ASAE
33:373–381.
Nelson, N. 1944. A photometric adaptation of the Somogyi method
for the determination of glucose. J. Biol. Chem. 153:375–380.
Treatment of 30 and 54% DM alfalfa with a single
strain of L. plantarum MTD1 improved the rapidity of
pH decline and quality of fermentation in laboratory
silos. At a DM content of 30%, there was no difference
in the decline in pH between the liquid and the dry
forms of inoculation. Similarly, in silage with the higher
DM, both forms of inoculant improved fermentation
relative to untreated silage. However, application of the
inoculant as a liquid resulted in faster growth of LAB,
increased production of lactic acid, and caused a faster
decline in silage pH than when the inoculant was ap-
plied as a dry formulation. The results of these studies
emphasize the fact that even under optimal ensiling
conditions, alfalfa can benefit from treatment with a
proven homolactic acid bacteria over a wide range of
DM. In addition, if forage is ensiled with a high DM
content, treatment with an inoculant applied in water
can be more effective than if it is applied in a dry-
granular form because the former process most likely
results in a better distribution and a better environ-
ment for a quicker reactivation of the bacteria. The
results of this study are specific to L. plantarum MTD1
and should not be extrapolated to other inoculants and Pahlow, G., and F. Weissbach. 1996. Effect of numbers of epiphytic
lactic acid bacteria (LAB) and of inoculation on the rate of pH-
their constituent strains of bacteria because different
decline in direct cut and wilted grass silages. Pages 104–105 in
strains have different tolerances to osmotic pressure.
Furthermore, there may be major differences in the rate
of reactivation of the various commercial inoculants.
Future studies should determine the relative osmotol-
erance of the microbes used in silage inoculants in order
to predict the likelihood that using such products would
be beneficial when applied to forage with high concen-
trations of DM.
Proc. XI International Silage Conference. Aberystwyth, Wales.
Pahlow, G., and F. Weissbach. 1999. New aspects of evaluation and
application of silage additives. Landbauforsch. Volkenrode
206(special issue):141–158.
Pitt, R. E., R. E. Muck, and R. Y. Leibensperger. 1985. A quantitative
model of the ensilage process in lactate silages. Grass Forage Sci.
40:279–303.
SAS User’s Guide: Statistics, Version 7 Edition. 1998. SAS Inst. Inc.,
Cary, NC.
Snedecor, G. W., and W. G. Cochran. 1980. Statistical Methods. 6th
ed. Iowa State Univ. Press, Ames, IA.
Stokes, M. R. 1992. Effects of an enzyme mixture, an inoculant, and
their interaction on silage fermentation and dairy production. J.
Dairy Sci. 75:764–773.
ACKNOWLEDGMENTS
Toller, J. A., and J.H.B. Christian. 1978. Water Activity and Food.
Academic Press. New York, NY.
Weatherburn, M. W. 1967. Phenol-hypochlorite reaction for determi-
nations of ammonia. Anal. Chem. 39:971–974.
Weissbach, F. 1968. Relationship between silage crop and cause of
fermentation during silage conservation. Ph.D. Thesis. Oskar-
Keller-Institute for Animal Nutrition. Rostook, Germany.
The authors thank the employees of the University
of Delaware Farm for assistance with harvesting of the
crops. C. M. Golt and J. D. Pesek are acknowledged
for help with VFA analysis and statistical analysis,
respectively. The authors also acknowledge the analyti-
Journal of Dairy Science Vol. 84, No. 10, 2001