Content uploaded by Giorgio Borreani
Author content
All content in this area was uploaded by Giorgio Borreani on Jan 14, 2016
Content may be subject to copyright.
Alfalfa
942 Agronomy Journal • Volume 100, Issue 4 • 2008
Published in Agron. J. 100:942–948 (2008).
doi:10.2134/agronj2007.0258
Copyright © 2008 by the American Society of Agronomy,
677 South Segoe Road, Madison, WI 53711. All rights
reserved. No part of this periodical may be reproduced
or transmitted in any form or by any means, electronic
or mechanical, including photocopying, recording, or
any information storage and retrieval system, without
permission in writing from the publisher.
B
has advantages over traditional
hay production, such as a more exible harvest date, less
weather dependency, and greater exibility in ration formula-
tion (Savoie and Jofriet, 2003). e bale silage technique is
characterized by its unique individual-package storage sys-
tem, but is particularly prone to spoilage, because each ton of
baled silage has six to eight times the surface area in contact
with plastic lm compared to clamp silage and about half of
the silage is within 120 mm of the lm cover (Forristall and
O’Kiely, 2005). Air penetration in the silage, which stimulates
aerobic bacteria, yeasts, and molds, is the main cause of aero-
bic deterioration; this results in DM and nutrient losses, the
accumulation of pathogens and mycotoxins (Scudamore and
Livesey, 1998), and reduced DM intake. erefore, to ensure
a good and stable silage conservation, air-tightness has to be
maintained throughout extended conservation periods (Paillat
and Gaillard, 2001).
Traditionally, manufacturers of plastic lms for bale wrap-
ping tended to use PE because of its mechanical characteristics
and low cost. Typical plastic lms for stretch-wrap silage are
coextruded, linear low-density PE that are 25 µm thick before
being stretched 50% or more during application (Lingvall,
1995). e low density of 0.92 g cm
−3
causes these PE lms to
be relatively permeable to oxygen and other gases (McNally et
al., 2005). Oxygen ow through a 25-µm-thick lm of a low-
density PE at 0.1 MPa overpressure, 23°C, and 85% relative
humidity is 7120 cm
3
m
−2
d
−1
(American Society for Testing
and Materials, 1980). Another important factor that is o en
neglected is the in uence of temperature on the permeability
coe cient that increases exponentially with temperature.
is leads to a 7.4-fold increase in permeability when lm
heats to 70°C in the sun (Daponte, 1994). Daponte (1992)
also reported that the oxygen permeabilities of PE silage lms
from low quality to high quality are close to each other in
the range of the thickness because a nearly linear relationship
exists between permeability and lm thickness. McNally et al.
(2005) showed that PE lms with low density have poorer gas
barrier properties than materials of higher density, crystallin-
ity, and orientation as a result of extrusion processing condi-
tions. Paillat and Gaillard (2001) reported that stretching to
60% reduced the thickness from 25 to 19 µm, accelerated lm
wear, and decreased the service life of the lm by 48% on aver-
age. Furthermore, Hancock and Collins (2006) reported that
the oxygen permeability of a single layer of PE lm stretched
to 150% of its original length increased to values ranging from
7750 to 9810 cm
3
m
−2
d
−1
, varying according to the manufac-
turing process.
Increasing the number of lm layers markedly improves air-
tightness because of the increase in distance the air has to travel
to reach the forage. Four layers of PE are usually applied in two
subsequent and complete rotations of the bale with an overlap
of 50% between the layers (Savoie and Jofriet, 2003). A signi -
cant reduction in mold growth and an improvement in silage
conservation quality were obtained when six or eight layers of
lm were applied compared to four (Keller et al., 1998; Müller,
2005). is is particularly true when high DM forages, and
especially alfalfa, are ensiled in wrapped bales and conserved
for periods more than 8 mo. More layers of stretch lm assure
a better airtight cover but involve prohibitive increases in costs,
ABSTRACT
e objectives of this research were to (i) test a new oxygen barrier (OB) stretch lm with a 20-fold lower oxygen permeability
than the polyethylene (PE) lm commonly used on farms to wrap bales and (ii) determine the e ects on microbial status, dry
matter (DM) losses, and fermentation of alfalfa (Medicago sativa L.) bale silage. Five eld trials were conducted on a farm near
Turin, Italy. In Trial 1, the bales were wrapped with two, four, six, or eight layers of either conventional PE or OB lm. A further
four trials were conducted to compare bales wrapped with four layers of OB with bales wrapped with six layers of PE. In the outer
layer of bales, the pH was lower in the OB than the PE silages. ere was a signi cant e ect of the lm type and number of lay-
ers on the percentage of bale surface covered by mold (P < 0.001), with less than 15% in bales wrapped with at least four layers of
OB lm. Storage DM losses ranged from 50 to 123 g kg
−1
DM and were a ected by the type of lm used (P < 0.001) and by the
number of layers applied (P = 0.035), with consistently lower values in OB silages. e new stretch lm supports the possibility
of conserving high quality alfalfa silage for up to 14 mo, with less lm than currently used.
New Oxygen Barrier Stretch Film Enhances Quality of Alfalfa
Wrapped Silage
Giorgio Borreani and Ernesto Tabacco*
Dip. di Agronomia, Selvicoltura e Gestione del Territorio, Univ. of Turin, via
Leonardo da Vinci, 44, 10095 Grugliasco (TO), Italy. Received 25 July 2007.
*Corresponding author (ernesto.tabacco@unito.it).
Abbreviations: cfu, colony forming units, CP, crude protein; DM, dry matter;
F, type of stretch lm e ect; HPLC, high performance liquid chromatography;
L, number of layers e ect; NH
3
–N, ammonia nitrogen; PE, polyethylene, OB,
oxygen barrier; TN, total nitrogen.
Agronomy Journal • Volume 100, Issue 4 • 2008 943
in plastic usage, and in environmental concern due to the dis-
posal of the additional plastic (Lingvall, 1995).
e rapid development in wrapping bale technology has led
to a great improvement in the ensiling process, by increasing
bale densities with round balers equipped with a crop-cutter
(Borreani and Tabacco, 2006), reducing working times with
combined baler-wrapper machines (Münster, 2001), and
improving uniformity of plastic distribution on the bale sur-
face with a new-concept wrapping system (Borreani et al.,
2007). However, PE has a permeability to oxygen that is too
high for this kind of application and, therefore, becomes the
weak link in the entire conservation process; a new plastic
material with low gas permeability is, therefore, required and
needs to be tested for baled silage.
e aim of this work was to test the e ects of two 25-µm
self-sealing stretch lms (conventional PE and a new OB lm)
and of an increasing number of plastic layers on the microbial
status, DM losses, and fermentation characteristics of baled
silage made from wilted alfalfa.
MATERIALS AND METHODS
Site and Experimental Design
Five trials were performed at the experimental farm of the
University of Turin in the western Po plain, northern Italy
(44°50´ N, 7°40´ E, altitude 232 m above sea level, annual
mean temperature 11.7°C, and annual average rainfall 739
mm). e research was conducted in 2005 and 2006 on a pro-
ducing 2.0-ha alfalfa eld at ve di erent harvests.
e rst trial (Trial 1) in 2005 was to study the e ects of the
type of stretch lm (conventional PE and OB) and the num-
ber of plastic layers applied (two, four, six, or eight layers) on
fermentation quality, mold development on bale surface, and
DM losses. Four further trials (Trials 2, 3, 4, and 5) were con-
ducted on a eld scale to support the hypothesis of reducing the
amount of plastic consumed without reducing the conservation
quality of bale silage, by using four layers of the new OB lm
instead of six layers of PE lm.
Trial 1
A second cut of alfalfa was mown with a rubber roll mower-
conditioner at 11:00 h. at a stubble height of 50 to 80 mm
and forage was spread by tedder over the whole eld surface
within 4 h. A er 1 d of eld wilting, the forage was raked
and baled ( xed chamber, no knives in the pickup, net wrap-
ping, Columbia R500/Z, WOLAGRI, Suzzara- MN, Italy) in
600-mm-long and 600-mm-diam. round bales and the silage
bales were individually wrapped (FW 500/Z, WOLAGRI,
Suzzara- MN, Italy) using two, four, six, or eight layers of
conventional PE stretch lm (white low density PE, 250 mm
wide × 25 m thick, Campanini Ugo S.p.A., Pieve di Cento-
BO, Italy, with an oxygen permeability at 1 bar overpressure,
23°C, and 85% relative humidity of 7160 cm
3
m
−2
d
−1
, 50%
stretched) or two, four, six, or eight layers of OB stretch lm
(transparent coextruded and without UV protection, 250 mm
wide × 25 m thick, speci cally produced for this experiment
by IPM S.p.A., Mondovì- CN, Italy, with an oxygen permeabil-
ity at 1 bar overpressure, 23°C, and 85% relative humidity of
400 cm
3
m
−2
d
−1
, 50% stretched). ree individual randomly
selected bales were wrapped for each treatment. e bales were
stored indoors in the dark on their ends in a single tier for over
9 mo (273 d of conservation).
Trials 2 to 5
First, second, third, and fourth cuts of alfalfa were mown
with a rubber roll mower-conditioner at a stubble height of
50 to 80 mm, and forage was spread by tedder over the whole
eld surface within 2 to 4 h. A er 1 to 3 d of eld wilting, the
forages were raked and baled ( xed chamber, no knives in the
pickup, net wrapping, Columbia R10, WOLAGRI, Suzzara-
MN, Italy) in 1200-mm-long and 1200-mm-diam round bales
and the silage bales were individually wrapped (Kverneland
Silawrap 7525, Kverneland Group, Naerbo, Norway) using six
layers of conventional PE stretch lm (white low density PE,
500 mm wide × 25 m thick, Campanini Ugo S.p.A., Pieve di
Cento- BO, Italy, 50% stretched) or four layers of OB stretch
lm (transparent coextruded and without UV protection, 500
mm wide × 25 m thick, speci cally produced for this experi-
ment by IPM S.p.A., Mondovì- CN, Italy, 50% stretched). Four
individual randomly selected bales were wrapped for each treat-
ment. e bales were stored outdoors on a concrete surface on
their side in a single tier for conservation periods of 331, 418,
122, or 167 d for respective trials. All the bales were covered by
a 200-m white plastic lm to avoid bird damage and to pro-
tect them from direct sunlight.
Forage Sampling and Silage Evaluation
e DM yield was estimated by weighing samples from
four randomly selected 1.8 × 4 m
2
areas at mowing time. Four
herbage subsamples were taken for DM and chemical analyses.
A er baling, two cores were taken from the side of the bale
from a depth of 0 to 480 mm with a core sampler (45 mm
diam.) and mixed together for DM concentration analysis.
A er the conservation period, before removing the plastic lm,
each bale was examined carefully for visible holes or damage
due to stem punctures. On removal of the plastic lms, all
visible molds on the bale surface were located and measured,
according to O’Brien et al. (2005). e percentage of the total
surface area a ected by fungal growth was then calculated for
each bale. Four cores were taken from the side of the bale from
a depth of 0 to 120 mm with a core sampler (45 mm diam.) and
mixed together, for microbiological analyses. Samples were also
taken from a depth of 121 to 480 mm from the bale surface
using the same holes, for microbiological and fermentative
analyses. e corer was disinfected between samples and bales
using 95% industrial methylated spirits. In addition, the outer
part of the bale (30 mm) was manually removed, weighed,
chopped, and subsampled for microbiological and fermentative
analyses, to better characterize the silage closer to the plastic
lm. e silage was chopped with a laboratory chopper (proto-
type CNR-IMAMOTER, Turin, Italy) set at a 50-mm theo-
retical length. e chopper knives were disinfected between
bales using 95% industrial methylated spirits. e samples for
microbiological analysis were immediately stored at 4°C before
analysis later in the day. Furthermore, in Trial 1, the dry weight
losses were evaluated by weighing the bales a er wrapping at 0,
36, 113, and 273 d of conservation.
944 Agronomy Journal • Volume 100, Issue 4 • 2008
Analytical Procedures
e DM concentration was determined on herbage samples
at 90°C dried in a forced-dra oven until constant weight.
Herbage subsamples were dried for chemical analyses by oven
drying to a constant weight at 60°C, then air equilibrated,
weighed, and ground in a Cyclotec mill (Tecator, Herndon,
VA) to pass a 1-mm screen. e dried samples were analyzed
for total N (TN) by combustion (Micro-N nitrogen analyser,
Elementar, Hanau, Germany) and for crude protein (CP) (TN
× 6.25). e silage samples were subsampled and immediately
analyzed for the DM concentration by oven drying at 80°C
for 24 h and for microbiological counts. Colony-forming units
(cfu) of yeasts and molds were counted using the pour plate
technique with 40.0 g L
−1
of yeast extract glucose chloram-
phenicol agar (DIFCO, West Molesey, Surrey, UK) a er incu-
bation at 25°C for 3 and 5 d for yeast and mold, respectively.
Clostridial spores were determined following the most prob-
able number technique with lactate-acetate agar (Spoelstra,
1984) a er incubation at 37°C for 7 d.
Wet silage samples (50 g) were homogenized and extracted
for 4 min in a Stomacher blender (Model 400, Seward Ltd,
London, UK) in water (220 mL) or in H
2
SO
4
0.1 N (220
mL). e pH was determined in the water extracts. Ammonia
nitrogen (NH
3
–N), determined using a speci c electrode,
was quanti ed in the water extracts. e lactic and volatile
fatty acids (acetic, butyric, propionic acids) were determined
by high performance liquid chromatography (HPLC) (Canale
et al., 1984). Ethanol was determined by HPLC, coupled to
a refractive index detector, on a Aminex HPX-87H column
(Bio-Rad Laboratories, Richmond, CA). Duplicate analyses
were performed for all the determined parameters. Duplicates
were averaged and the means (replicated bales) were considered
as observations in the statistical analysis.
Statistical Analysis
e data were analyzed for their statistical signi cance by
ANOVA, with their signi cance reported at a 0.05 probability
level using the general linear model of the Statistical Package
for Social Science (v 11.5, SPSS Inc., Chicago, IL). In Trial 1,
data were analyzed by ANOVA utilizing the type of stretch
l m e ect (F), the number of plastic layers e ect (L), and the F
× L interaction as xed factors with three replicates. In Trials
2 to 5, signi cant di erences between means were identi ed by
the P values of ANOVA and e ects were considered signi cant
at P < 0.05. e mold surface coverage data, expressed as a bale
surface percentage, were analyzed as angular transformed val-
ues (arcsine transformation). e data of DM losses observed
in Trial 1 were regressed with mold surface coverage as the
independent variable. e MANOVA analysis of covariance
was used to verify the equivalence of the equations for the
lm types. e regression lines related to each lm type were
pooled because they were not signi cantly di erent. e data
of mold surface coverage observed in Trials 1 to 5 were aver-
aged across eld replicates and regressed on the number of days
of conservation as the independent variable. e determination
coe cient (r
2
), adjusted for degrees of freedom, and RMSE are
reported.
RESULTS AND DISCUSSION
e DM yields, the DM and the CP concentrations at cut-
ting and at baling, the hours of wilting, and the days of conser-
vation are reported in Table 1. Weather was generally favorable
for eld drying during each trial and no alfalfa was rained
on when drying in the eld. e favorable drying conditions
allowed the alfalfa to be harvested a er 27, 48, 56, 72, and 32
h of wilting, for the respective trials. Collectively, these tri-
als encompassed a wide range of climatic conditions resulting
in various wilting periods, DM content at baling, and silage
conservation periods. e DM yields and the DM and CP
concentrations at cutting were typical of the alfalfa in the Po
plain (Tabacco et al., 2002). e decrease in CP concentration
was due to mechanical losses that occur during wilting and
was in the range of values observed in the same environment
by Borreani and Tabacco (2006) for alfalfa baled at a similar
DM content. e DM concentration at baling ranged from 459
to 720 g kg
−1
. Concentrating as much DM per bale as possible
was desirable in terms of plastic consumption and costs, bale
weight, and bale number per hectare (Beaulieu et al., 1993).
Ensiling alfalfa with more than 450 g DM kg
−1
, however,
makes silage more prone to spoilage and may result in molding
and heating, which is apparently related to di culties in the
exclusion of oxygen (Han et al., 2006).
Trial 1
e bale weight ranged from 36 to 48 kg bale
−1
, with an aver-
age value of 42 kg. e bale DM density ranged from 131 to 176
kg m
−3
, with an average value of 155 kg m
−3
. ese values were
similar to the DM densities of 157 kg m
−3
reported by Beaulieu
et al. (1993) for round bale silage with
greater dimensions (1200 mm diam. and
1200 mm long), and were consistent with
those reported by Huhnke et al. (1997) in
the range of 101 to 231 kg DM m
−3
.
e chemical composition of alfalfa
bale silages a er 273 d of conservation is
given in Table 2. e forage DM concen-
trations were relatively uniform across
the eld and no di erences in silage DM
concentration were observed among
treatments at harvesting. In the present
study, the lactic acid concentrations and
ammonia were not a ected by the lm
type or the number of layers (P > 0.05)
Table 1. Harvest date, main characteristics of alfalfa herbage at cutting and at baling,
bale weight ,and bale DM density of the ensiling trials conducted at Turin, Italy.†
Trial
12 345
Cut second fi rst fourth second third
Cutting date 23 June 2005 11 May 2006 23 Sept. 2005 2 July 2006 16 Aug. 2006
Wilting time, h 27 48 56 72 32
DM yield, Mg ha
–1
3.9 4.8 1.9 3.3 2.9
DM concentration at cutting, g kg
–1
214 202 239 218 207
CP at cutting, g kg
–1
DM 192 206 199 193 195
DM concentration at baling, g kg
–1
642 459 720 554 677
CP at baling, g kg
–1
DM 164 194 156 153 160
Days of conservation, d 273 331 418 122 167
† DM, dry matter; CP, crude protein.
Agronomy Journal • Volume 100, Issue 4 • 2008 945
and were in the range of values reported by Huhnke et al.
(1997) for legume bale silages of similar DM content wrapped
with at least six layers of PE and analyzed a er at least 6 mo
in storage. e high concentration of DM likely restricted
fermentation and resulted in relatively low concentrations of
all fermentation acids and NH
3
–N. is was consistent with
observations reported for wilted and severely wilted ( > 600
g DM kg
−1
) alfalfa round bale silages by Han et al. (2006);
these researchers suggested that a high DM concentration
depresses the total amount of fermentation in silages, resulting
in a higher nal pH and lower concentrations of fermentation
acids, particularly lactic acid. ere was a signi cant lm type
e ect on pH and acetic acid for both the surface samples and
the core samples, with lower pH and greater acetic acid concen-
trations in silage wrapped with OB lm. e number of layers
in uenced the pH of the surface samples, with values decreas-
ing with increasing amount of plastic applied. e pH was in
the range of values reported for legume-grass bale silage wilted
to a higher DM concentration than 600 g kg
−1
(Huhnke et
al., 1997), except for the outer layer (0–30 mm) of the silages
wrapped with two layers of PE lm, where a pH of 6.45 indi-
cates evidence of silage spoilage, as suggested by Huhnke et al.
(1997). ere was a signi cant lm type and number of lay-
ers e ect on the ethanol concentration measured in the outer
layer of the bale, with lower values in silage wrapped with the
OB lm and decreasing values with an increasing number of
applied layers. A greater level of ethanol in the outer layer of the
silages wrapped with the PE lm indicated a possibly greater
permeation of oxygen through the cover. McDonald et al.
(1991) pointed out that the survival of aerobic microbes, whose
presence promotes maximum degradation, is shown by an
increased silage temperature and a production of compounds
such as ethanol and 2,3 butanediol. For the PE bales, ethanol
is signi cantly higher with four layers and increases dramati-
cally from two to four layers. It can be supposed that in bales
wrapped with two layers of PE, ethanol was partially oxidized
as suggested by Spoelstra et al. (1988) who reported that in
maize silage exposed to air, ethanol was oxidized by acetic acid
bacteria to acetic acid followed by a rapid oxidation of lactic
and acetic acids when ethanol was depleted. As expected, no
detectable butyric acid was present in these high DM silages.
High butyric acid production is normally associated with
undesirable clostridial fermentations that usually occur when
concentrations of DM are lower than 300 g kg
−1
(McDonald et
al., 1991).
Some holes were observed in the plastic lm for the bales
wrapped with two or four layers but there was no di er-
ence between the types of stretch lm (Table 3). e average
numbers of holes were 6 and 1 holes per bale, for the two and
four layers, respectively, and they were due to puncturing of
the lm by sti alfalfa stems. ere were signi cant lm type
and number of layer e ects on the percentage of bale surface
covered by mold (P < 0.001), with values decreasing when
the number of applied layers increased. e OB lm greatly
reduced the area of the surface of the bale covered by mold,
Table 3. Number of holes in the plastic, percentage of bale sur-
face covered by visible mold, and DM losses in relation to the
type of stretch fi lm and number of layers on second cut of alfal-
fa bale silage after 273 d of conservation at Turin, Italy, Trial 1.†
Stretch fi lm Layers
Bales
with
holes
Holes
per
bale
Bale surface covered
by mold
DM
lossesEnds Side Total
% g kg
–1
PE 2 3 6.3 100.0 82.5 94.2 123
4 1 0.7 86.7 73.3 82.2 95
6 0 0.0 41.2 30.0 37.5 86
8 0 0.0 37.0 35.0 36.4 80
OB 2 3 6.7 25.5 55.8 35.6 75
4 1 1.0 2.5 18.2 7.7 51
6 0 0.0 0.7 5.8 2.4 51
8 0 0.0 0.3 2.3 1.0 50
F (P value)‡ – <0.001 <0.001 <0.001 <0.001
L (P value) ‡ – <0.001 <0.001 <0.001 0.035
F × L (P value)‡ – 0.615 0.002 0.073 0.863
SED§ 0.20 0.11 0.10 19.9
† DM, dry matter; F, type of stretch fi lm effect; L, effect of the number of layers;
OB, oxygen barrier stretch fi lm; PE, polyethylene stretch fi lm.
‡ Effects were considered signifi cant at P < 0.05.
§ For mold surface coverage the SED is reported as an angular transformed value.
Table 2. Fermentative profi le in relation to the type of stretch fi lm and number of layers on second cut of alfalfa bale silage after
273 d of conservation at Turin, Italy, Trial 1.†
Stretch fi lm Layers
Bale surface (0–30 mm) Bale core (121–480 mm)
DM pH
Lactic
acid
Acetic
acid Ethanol NH
3
–N DM pH
Lactic
acid
Acetic
acid Ethanol NH
3
–N
g kg
–1
g kg
–1
DM g kg
–1
TN g kg
–1
DM g kg
–1
TN
PE 2 664 6.45 1.02 0.53 0.91 24.4 643 5.68 1.26 0.60 1.04 38.5
4 614 5.85 4.02 1.38 6.60 22.4 629 5.63 0.70 0.85 5.46 27.9
6 667 5.80 1.12 0.59 4.54 23.0 693 5.68 0.32 0.48 3.23 21.1
8 681 5.74 0.51 0.39 3.30 15.5 670 5.66 0.54 0.60 3.75 19.1
OB 2 577 5.73 3.71 2.16 2.89 23.1 579 5.63 3.25 2.49 3.40 31.4
4 620 5.59 1.57 1.42 3.06 19.3 642 5.56 0.55 0.84 2.82 17.7
6 634 5.58 1.02 1.03 2.12 15.1 603 5.63 0.64 1.00 2.61 25.3
8 645 5.65 1.25 1.40 1.13 23.2 634 5.62 1.44 1.25 1.76 25.6
F (P value)‡ 0.070 0.001 0.719 0.004 0.023 0.580 0.062 0.026 0.218 0.026 0.136 0.656
L (P value) ‡ 0.289 0.019 0.106 0.220 0.015 0.141 0.781 0.130 0.180 0.291 0.056 0.077
F × L (P value) ‡ 0.430 0.189 0.059 0.132 0.028 0.188 0.317 0.114 0.626 0.212 0.007 0.297
SED 47.6 0.48 1.49 0.58 1.50 6.88 49.7 0.08 1.46 0.76 1.13 9.28
† DM, dry matter; F, type of stretch fi lm effect; L, number of layers effect; NH
3
–N, ammonia nitrogen; OB, oxygen barrier stretch fi lm; PE, polyethylene stretch fi lm.
‡ Effects were considered signifi cant at P < 0.05.
946 Agronomy Journal • Volume 100, Issue 4 • 2008
with lower values than 8% even in the bales wrapped with
only four layers. e silage wrapped with two layers of OB lm
presented the same amount of molded surface as bale silages
wrapped with six or eight layers of PE, whereas the surface of
silages wrapped with two or four layers of PE had more than
80% of the surface covered by molds. e greater percentage of
bale surface covered by mold was re ected by the higher mold
count observed in the outer layer of silage wrapped with the PE
lm in comparison to the OB lm, independent of the number
of layers applied (Table 4). e outer 30-mm part of the bale
accounted for about 20% of the total DM stored in the bale
(data not shown), meaning that 20% of the silage wrapped with
two or four layers of PE had to be discarded before feeding to
animals. e di erences were reduced when a deeper layer of
the bale was considered (0 to 120 mm from the surface). e
mold count was signi cantly higher only in silages wrapped
with two layers of PE, meaning that more than 50% of the DM
stored in the bale was a ected by spoilage. No di erences in
mold count were observed for the silage sampled in the core of
the bale. e mold counts in the outer 30-mm layer of the bales
wrapped with PE con rmed the values reported by O’Brien et
al. (2007), who found 4.75 log mold cfu g
−1
in bale silages with
intact lm and wrapped with four layers of plastic, and 5.18
log mold cfu g
−1
in bale silages with visible damage to their PE
wrapping. Our results are consistent with the data reported
by Keller et al. (1998), who found values of 4.0 to 5.5 log mold
cfu g
−1
when less than six layers of PE lm were used to wrap
alfalfa silages with a DM concentration ranging from 331 to
547 g kg
−1
.
e DM losses at the end of the storage period ranged from
50 to 123 g kg
−1
DM (Table 3) and were a ected by the type
of stretch lm used (P < 0.001) and by the number of layers
applied (P = 0.035). Hancock and Collins (2006) reported
that DM losses for alfalfa round bale silage at a DM content of
626 g kg
−1
were generally low but highly variable, averaging 63
± 64 g kg
−1
DM, independent of the amount of PE stretch lm
applied (from two to six layers per bale). e pattern of weight
losses during the conservation period di ered from the rst
36 d of conservation (Fig. 1). e bales wrapped with at least four
layers of OB lm showed similar trends in weight loss, which were
consistently lower than those observed in silage wrapped with two
layers of OB lm or in all silages wrapped with the PE lm. A er
120 d of conservation, only the bales wrapped with two layers of
stretch lm (either OB or PE) showed limited increases in weight
losses, whereas the other bales did not show any further weight
losses. e regression equation of the pooled data of DM losses (n
= 24) on the percentage of the bale surface covered by mold are
reported in Fig. 2, where a positive relation can be seen between
the surface covered by molds and the nal DM losses, with an
adjusted coe cient of determination of 0.66 and a RMSE of 17.4.
Table 4. Mold and yeast count, clostridial spores (MPN), and pH in three zones of the bale in relation to the
type of stretch fi lm and number of layers on second cut of alfalfa bale silage after 273 d of conservation at
Turin, Italy, Trial 1.†
Stretch fi lm Layers
Bale surface (0–30 mm) Bale surface (0–120 mm) Bale core (121–480 mm)
pH Mold Yeast Spores pH Mold Yeast‡ Spores pH Mold Yeast‡ Spores
–log cfu g
–1
– MPN g
–1
–log cfu g
–1
– MPN g
–1
–log cfu g
–1
– MPN g
–1
PE 2 6.45 4.8 3.7 1.2 5.89 4.1 <1.0 1.0 5.68 2.4 <1.0 1.3
4 5.85 4.4 3.7 1.9 5.70 2.8 1.9 1.0 5.63 1.0 <1.0 1.2
6 5.80 4.2 3.7 1.5 5.79 1.8 <1.0 1.2 5.68 2.2 <1.0 1.5
8 5.74 4.4 4.5 2.2 5.70 2.6 3.3 1.3 5.66 1.7 <1.0 1.2
OB 2 5.73 4.2 3.7 1.0 5.53 1.9 <1.0 1.0 5.63 1.9 <1.0 1.6
4 5.59 3.7 3.6 1.5 5.67 2.3 <1.0 1.4 5.56 1.9 <1.0 1.0
6 5.58 3.7 4.1 1.3 5.58 2.0 <1.0 1.5 5.63 2.2 <1.0 1.5
8 5.65 3.4 3.9 1.3 5.58 2.2 <1.0 1.0 5.62 2.5 <1.0 1.2
F (P value)§ 0.001 <0.001 0.595 0.049 <0.001 0.007 – 0.304 0.026 0.103 – 0.771
L (P value)§ 0.019 0.003 0.020 0.134 0.650 0.044 – 0.077 0.130 0.014 – 0.186
F × L (P value)§ 0.189 0.369 0.069 0.566 0.054 0.021 – 0.127 0.114 0.028 – 0.790
SED 0.48 0.89 0.92 0.60 0.17 1.42 0.23 0.08 0.99 0.47
† cfu, colony forming units; F, type of stretch fi lm effect; L, effect of the number of layers; MPN, most probable number; OB, oxygen barrier
stretch fi lm; PE, polyethylene stretch fi lm.
‡ Statistic analysis was not performed.
§ Effects were considered signifi cant at P < 0.05.
Fig. 1. Dry weight losses during conservation in relation to
the type of stretch film and number of layers on second cut
of alfalfa bale silage in Trial 1 at Turin, Italy. PE2, PE4, PE6,
and PE8 indicate polyethylene stretch film of two, four, six, or
eight layers, respectively; OB2, OB4, OB6, and OB8 indicate
oxygen barrier stretch film of two, four, six, or eight layers,
respectively. Each symbol corresponds to the average of four
replicates. SEM = 2.70.
Agronomy Journal • Volume 100, Issue 4 • 2008 947
Trials 2 to 5
ese further four trials were performed at a farm scale to
verify the hypothesis of using four layers of OB lm instead
of six layers of PE lm to reduce plastic consumption without
increasing the risk of mold damage.
No holes were observed in either plastic lm a er the con-
servation period. All silages were well fermented and no di er-
ences were found between the fermentative pro les of the core
samples for the two treatments (Table 5). At high DM con-
centration, the fermentation was restricted and resulted in low
concentrations of all the fermentation products. e NH
3
–N
concentrations were signi cantly di erent in Trial 3, where the
silage wrapped with PE lm and analyzed a er 418 d of conser-
vation had an NH
3
–N value of 189 g kg
−1
TN in comparison
with the 49 g kg
−1
of the OB silage. When the outer 30-mm
layer of the bale silage was analyzed, signi cant di erences were
found between the two treatments in all the trials, in terms
of pH, mold and yeast counts, and
surface area covered by mold (Fig.
3). e pH was numerically higher
in the PE silages in all the trials
with values signi cantly higher in
Trials 3, 4, and 5. Yeast counts were
signi cantly higher in PE silages
than OB silages in Trials 2, 3, and
4, whereas it was lower in PE silages
in Trial 5. Mold counts were signi -
cantly higher in PE silages than in
OB silages in Trials 2 and 4. Yeast
and mold counts were slightly lower
than values observed in Trial 1, and
values higher than 3 log cfu g
−1
were
only observed in PE silages of Trial
2 and 4. e surface area of the bale
covered by mold was only higher
than 15% of the total surface in the
PE silages in Trials 2 and 3, which
were conserved for more than 300 d.
is suggested analyzing the pooled
data of the ve trials by regressing
the surface area covered by mold against the number of days of
conservation (Fig. 4). e percentage of the surface covered by
mold increased linearly with an increasing number of days of
conservation, both for PE and OB silages, with regressions with
high adjusted coe cient of determination (0.68 and 0.86 for
OB and PE, respectively) and low RMSE. e increase in surface
covered by mold per day of conservation was signi cantly higher
in PE silages than in OB silages and a er 300 d of storage more
than 30% of the surface of the bales wrapped with six layers of
PE was covered by mold, indicating a high degree of spoilage. e
surface covered by mold in the OB silages did not exceed 15%
even a er 418 d of conservation.
e reduction in the oxygen permeability obtained with
four layers of the OB lm can improve the DM recovered with
Table 5. Fermentative profi les of the core (121–480 mm) of
alfalfa bale silages in trials 2 to 5 at Turin, Italy.†
Stretch
fi lm DM pH
Lactic
acid
Acetic
acid
Butyric
acid‡ NH
3
–N
g kg
–1
DM g kg
–1
TN
Trial 2 PE-6 467 5.44 11.2 3.5 0.0 90
OB-4 452 5.28 11.7 4.5 0.0 73
P value§ 0.240 0.106 0.934 0.499 – 0.112
Trial 3 PE-6 737 5.22 8.8 0.6 0.0 189
OB-4 703 5.37 7.0 0.2 0.0 49
P value§ 0.142 0.351 0.162 0.169 – 0.007
Trial 4 PE-6 546 5.27 12.1 9.0 0.6 80
OB-4 562 5.34 9.9 4.9 0.0 55
P value§ 0.375 0.247 0.385 0.277 – 0.652
Trial 5 PE-6 664 5.67 8.2 2.6 0.3 85
OB-4 690 5.79 11.3 4.7 0.0 91
P value§ 0.301 0.246 0.528 0.694 – 0.309
† DM, dry matter; NH
3
–N, ammonia nitrogen; PE- 6, bales wrapped with six lay-
ers of polyethylene stretch fi lm; OB- 4, bales wrapped with four layers of oxygen
barrier stretch fi lm.
‡ Statistic analysis was not performed.
§ Effects were considered signifi cant at P < 0.05.
Fig. 2. Dry matter (DM) losses at the end of conservation in
relation to bale surface covered by mold on the second cut
of alfalfa bale silage in Trial 1 at Turin, Italy. PE, polyethylene
stretch film; OB, oxygen barrier stretch film. Regression
equation of the pooled data: DM losses = 0.714 surface molds +
49.60; adjusted r
2
= 0.66; RMSE = 17.43.
Fig. 3. The pH, yeast and mold count in the outer 30 mm of the bale, and surface covered by
mold in alfalfa bale silages in Trials 2 to 5 at Turin, Italy. PE, 6 layers, bales wrapped with six
layers of polyethylene stretch film; OB, 4 layers, bales wrapped with four layers of oxygen
barrier stretch film. Effects were considered significant at P < 0.05.
948 Agronomy Journal • Volume 100, Issue 4 • 2008
the wrapped bale system, by reducing the mold development in
the silage layer closest to the lm and, consequently, reducing the
weight losses during conservation. e results obtained with four
layers of the OB lm consistently support the possibility of reduc-
ing the amount of plastic applied per ton of stored DM in severely
wilted forages and of improving the conservation and micro-
biological quality of silages that can only be obtained with at
least six layers of PE lm, as suggested by Lingvall (1995) and
Keller et al. (1998).
CONCLUSION
e new OB stretch lm reduced DM losses and mold spoil-
age in high DM alfalfa silages in comparison to the PE lm that
is typically used commercially. e data showed that the bale
surface covered by molds can be kept below 15% with four layers
of OB lm, whereas it reached up to 80% with the same amount
of PE lm. is supports the possibility of making good quality
alfalfa wrapped silage with four layers of plastic instead of the six
or even eight layers commonly suggested with PE lms for longer
conservation periods than 8 mo. is new tool may solve prob-
lems that have limited the application of wrapping technology
to extremely wilted alfalfa silage without increasing the amount
of plastic applied, which results in increased costs and environ-
mental concerns. Further experiments should be conducted to
investigate the e ects of the OB lm on di erent forage crops
and DM concentrations at ensiling.
ACKNOWLEDGMENTS
The authors thank Bartolomeo Forzano (Industria Plastica Monregalese
SpA of Mondovì, Italy) for providing the coextruded barrier plastic films
utilized in the experiment. The authors also like to thank Serenella
Piano, Mara Scaiola, and Matteo Maurelli (Dip. Agronomia, Selvicoltura
e Gestione del Territorio, Univ. of Turin, Italy) for the chemical and
microbiological analyses. This work was partially funded by the Regione
Lombardia, Direzione Generale Agricoltura, Project “MARINSIL.” The
authors contributed equally to the work described in this paper. Mention
of trade names is for the benefit of the reader and does not constitute
endorsement by the Univ. of Turin over other products not mentioned.
REFERENCES
American Society for Testing and Materials. 1980. ASTM Standard method
D 3985–81. In Annual Book of Standards. ASTM, Philadelphia, PA.
Beaulieu, R., J.R. Seoane, P. Savoie, D. Tremblay, G.F. Tremblay, an R.
ériault. 1993. Eff ects of dry-matter content on the nutritive value of
individually wrapped round-bale timothy silage fed to sheep. Can. J.
Anim. Sci. 73:343–354.
Borreani, G., C. Bisaglia, and E. Tabacco. 2007. Eff ects of a new-con-
cept wrapping system on alfalfa round bale silage. Trans. ASABE
50:781–787.
Borreani, G., and E. Tabacco. 2006. e eff ect of a baler chopping system on
fermentation and losses of wrapped big bales of alfalfa. Agron. J. 98:1–7.
Canale, A., M.E. Valente, and A. Ciotti. 1984. Determination of volatile
carboxylic acids (C
1
–C
5i
) and lactic acid in aqueous acid extracts of
silage by high performance liquid chromatography. J. Sci. Food Agric.
35:1178–1182.
Daponte, T. 1992. Coextruded fi lms in silage. Plasticulture 96(4):35–44.
Daponte, T. 1994. Barrier fi lms for soil fumigation. Plasticulture 102(2):17–24.
Forristall, P.D., and P. O’Kiely. 2005. Update on technologies for producing
and feeding silage. p. 83–96. In R.S. Park et al. (ed.) Proc. 14th Int.
Silage Conf., Belfast, Northern Ireland, 3–6 July 2005. Wageningen
Academic Publ., Wageningen, e Netherlands.
Han, K.J., M. Collins, E.S. Vanzant, and C.T. Dougherty. 2006.
Characteristics of baled silage made from fi rst and second harvest of
wilted and severely wilted forages. Grass Forage Sci. 61:22–31.
Hancock, D.W., and M. Collins. 2006. Forage preservation method infl u-
ences alfalfa nutritive value and feeding characteristics. Crop Sci.
46:688–694.
Huhnke, R.L., R.E. Muck, and M.E. Payton. 1997. Round bale silage stor-
age losses of ryegrass and legume-grass forages. Appl. Eng. Agric.
13:451–457.
Keller, T., H. Nonn, and H. Jeroch. 1998. e eff ect of sealing and of additives
on the fermentation characteristics and mould and yeast counts in stretch
fi lm wrapped big-bale lucerne silage. Arch. Anim. Nutr. 51:63–75.
Lingvall, P. 1995. e balewrapping handbook. Trioplast, AB,
Smalandsstenar, Sweden.
McDonald, P., A.R. Henderson, and S.J.E. Heron. 1991. e biochemistry
of silage. 2nd ed. Chalcombe Publ., Bucks, UK.
McNally, G.M., C. Laffi n, P.D. Forristal, P. O’Kiely, and C.M. Small. 2005.
e eff ect of extrusion conditions and material properties on the gas
permeation properties of LDPE/LLDPE silage wrap fi lms. J. Plast.
Film Sheeting 21:27–37.
Müller, C.E. 2005. Fermentation patterns of small-bale silage and haylage
produced as a feed for horses. Grass Forage Sci. 60:109–118.
Münster, J.M. 2001. Trends in forage harvesting technology. Landtechnik
56:386–387.
O’Brien, M., P. O’Kiely, P.D. Forristal, and H.T. Fuller. 2005. Fungi isolated
from contaminated baled grass silage on farms in the Irish Midlands.
FEMS Microbiol. Lett. 247:131–135.
O’Brien, M., P. O’Kiely, P.D. Forristal, and H.T. Fuller. 2007. Quantifi cation
and identifi cation of fungal propagules in well-managed baled grass
silage and in normal on-farm produced bales. Anim. Feed Sci. Technol.
132:283–297.
Paillat, J.M., and F. Gaillard. 2001. Air-tightness of wrapped bales and resis-
tance of polythene stretch fi lm under tropical and temperate condi-
tions. J. Agric. Eng. Res. 79:15–22.
Savoie, P., and J.C. Jofriet. 2003. Silage storage. p. 405–467. In D.R. Buxton
et al. (ed.) Silage science and technology. Agron. Monogr. 42. ASA,
CSSA, and SSSA, Madison, WI.
Scudamore, K.A., and C.T. Livesey. 1998. Occurrence and signifi cance of myco-
toxins in forage crops and silage: A review. J. Sci. Food Agric. 77:1–17.
Spoelstra, S.F. 1984. Some methods to evaluate the role of clostridia in silage.
Int. Rep. No. 168. IVVO-DLO, Lelystad, the Netherlands.
Spoelstra, S.F., M.G. Courtin, and J.A.C. van Beers. 1988. Acetic acid bac-
teria can initiate aerobic deterioration of whole crop maize silage. J.
Agric. Sci. 111:127–132.
Tabacco, E., G. Borreani, M. Odoardi, and A. Reyneri. 2002. Eff ect of cut-
ting frequency on dry matter yield and quality of lucerne (Medicago
sativa L.) in the Po Valley. Italian J. Agron. 6:27–33.
Fig. 4. Surface covered by mold (%) in relation to days of con-
servation in alfalfa bale silages in Trials 1 to 5 at Turin, Italy.
PE - 6 layers, bales wrapped with six layers of polyethylene
stretch film; OB - 4 layers, bales wrapped with four layers
of oxygen barrier stretch film; regression equations of the
pooled data of the five trials: Surface molds
PE-6
(%) = 0.183
× days of conservation – 20.5; adjusted r
2
= 0.86; RMSE = 8.9.
Surface molds
OB-4
(%) = 0.0442 × days of conservation – 6.36;
adjusted r
2
= 0.68; RMSE = 3.4.