Methane emissions from beef cattle: Effects of monensin, sunflower oil,
enzymes, yeast, and fumaric acid1
S. M. McGinn2, K. A. Beauchemin, T. Coates, and D. Colombatto3
Agriculture and Agri-Food Canada, Research Centre, Lethbridge, AB, Canada
ABSTRACT: Methane emitted from the livestock
sector contributes to greenhouse gas (GHG) emissions.
Understanding the effects of diet on enteric methane
production can help refine GHG emission inventories
and identify viable GHG reduction strategies. Our
study focused on measuring methane and carbon dioxide emissions, total-tract digestibility, and ruminal fermentation in growing beef cattle fed a diet supplemented with various additives or ingredients. Two experiments, each designed as a 4 × 4 Latin square with
21-d periods, were conducted using 16 Holstein steers
(initial BW 311.6 ± 12.3 kg). In Exp. 1, treatments were
control (no additive), monensin (Rumensin, Elanco Animal Health, Indianapolis, IN; 33 mg/kg DM), sunflower
oil (400 g/d, approximately 5% of DMI), and proteolytic
enzyme (Protex 6-L, Genencor Int., Inc., CA; 1 mL/kg
DM). In Exp. 2, treatments were control (no additive),
Procreatin-7 yeast (Prince Agri Products, Inc., Quincy,
IL; 4 g/d), Levucell SC yeast (Lallemand, Inc., Rexdale,
Ontario, Canada; 1 g/d), and fumaric acid (Bartek Ingredients Inc., Stoney Creek, Ontario, Canada; 80 g/d).
The basal diet consisted of 75% barley silage, 19%
steam-rolled barley grain, and 6% supplement (DM basis). Four large chambers (two animals per chamber)
were equipped with lasers and infrared gas analyzers
to measure methane and carbon dioxide, respectively,
for 3 d each period. Total-tract digestibility was determined using chromic oxide. Approximately 6.5% of the
GE consumed was lost in the form of methane emissions
from animals fed the control diet. In Exp. 1, sunflower
oil decreased methane emissions by 22% (P = 0.001)
compared with the control, whereas monensin (P = 0.44)
and enzyme had no effect (P = 0.82). However, oil decreased (P = 0.03) the total-tract digestibility of NDF
by 20%. When CH4 emissions were corrected for differences in energy intake, the loss of GE to methane was
decreased by 21% (P = 0.002) using oil and by 9% (P =
0.09) using monensin. In Exp. 2, Procreatin-7 yeast (P =
0.72), Levucell SC yeast (P = 0.28), and fumaric acid
(P = 0.21) had no effect on methane emissions, although
emissions as a percentage of GE intake were 3% (nonsignificant, P = 0.39) less for steers fed Procreatin-7
yeast compared with the control. This study demonstrates that sunflower oil, ionophores, and possibly
some yeast products can be used to decrease the GE
lost as methane from cattle, but fiber digestibility is
impaired with oil supplementation.
Key Words: Beef Cattle, Carbon Dioxide, Feed Additives, Greenhouse Gases, Methane
2004 American Society of Animal Science. All rights reserved.
Introduction
As evidence for global warming becomes prevalent,
there is growing consensus that the emissions of green-
1
Lethbridge Research Centre Contribution No. (387) 03100. We
thank our technicians, K. Andrews (animal care and sampling), L.
Kremenik (laboratory analysis), B. Baker (animal care), D. Steacy
(sample preparation), R. Wuerfel (laboratory analysis), and A. Furtado (laboratory analysis) for their significant contributions. We also
thank Genencor Int., Inc., for supplying the enzyme product, Prince
Agri Products, Inc., and Lallemand, Inc., for supplying the yeast
products, and Bartek Ingredients, Inc., for supplying the fumaric acid.
The Canadian federal government’s Model Farm Program funded this
study.
2
Correspondence: Box 3000, 5403 1st Ave. South (phone: 403-3274561; fax: 403-317-2182; e-mail: [email protected]; beauchemin@
agr.gc.ca).
J. Anim. Sci. 2004. 82:3346–3356
house gases (GHG) into our atmosphere must be mitigated. In 2000, the Canadian agricultural industry released 60.5 Mt of CO2 equivalent, accounting for 8.3%
of the GHG emitted from all Canadian sectors. Within
the agricultural sector, CH4 from livestock accounted
for 38% of the GHG emitted (Environment Canada,
2002). For the cattle industry, decreasing CH4 losses
can represent an improvement in feed efficiency. Thus,
mitigating CH4 losses from cattle has both long-term
environmental and short-term economic benefits.
3
Present address: Departamento de Producción Animal, Facultad
de Agronomı́a, Universidad de Buenos Aires, Av. San Martı́n 4453
(C1417DSQ), Buenos Aires, Argentina (e-mail: colombat@
agro.uba.ar).
Received January 15, 2004.
Accepted July 7, 2004.
3346
Methane emissions from beef cattle
Diet modifications can help mitigate CH4 emissions
from cattle. Dietary manipulations reduce CH4 emissions by decreasing the fermentation of OM in the rumen, shifting the site of digestion from the rumen to
the intestines, diverting H away from CH4 production
during ruminal fermentation, or by inhibiting methanogenesis by ruminal bacteria (Johnson and Johnson,
1995; Benchaar et al., 2001).
Ionophores have been shown to decrease CH4 emissions from cattle, although effects may be transient
(Johnson and Johnson, 1995). Furthermore, continued
future use of antimicrobials in animal production is
tenuous. Several alternative feed additives have been
recently shown in vitro to reduce CH4 emissions, including organic acids (Asanuma et al., 1999), enzymes (Colombatto et al., 2003a), and yeast (Sacchararomyces
cerevisiae; Mutsvangwa et al., 1992). However, there is
limited information to demonstrate the effects of these
additives in vivo. Some chemical compounds, such as
methane inhibitors, have been demonstrated to reduce
CH4 (Garcia- López et al., 1996; Miller and Wolin, 2001;
Itabashi, 2002); however, these compounds are not approved in Canada.
The purpose of our study was to investigate the effect
of several ingredients and feed additives that are currently registered for feeding to cattle on enteric CH4
production.
Experimental Procedures
This paper reports on two experiments conducted between September and December 2002, using the Controlled Environment Facility at Agriculture and AgriFood Canada Research Center in Lethbridge, Alberta.
All animals were cared for in accordance with the guidelines of the Canadian Council on Animal Care (1993).
Experimental Design
The two experiments were each designed as a 4 × 4
Latin square. Each experiment used eight cattle fed
four dietary treatments during four 21-d periods. The
experiments were offset by 1 wk to facilitate measurements. During the first 16 d of each period, the cattle
were housed in individual pens (4.9 × 1.8 m) bedded
with straw in a sheltered barn. Daily feed intake and
ruminal fermentation were measured during this
phase. Before the morning feeding on d 17, the cattle
were moved to four chambers for measurements of CH4,
CO2, and total-tract digestibility. Two animals were
housed in each chamber. The cattle were paired at the
start of the experiment such that the total weight of
cattle per chamber was similar. The pairing of animals
was consistent throughout the experiment, such that
animals within a chamber received the same treatment.
The first day within the chambers was considered an
adjustment period, allowing the steers to adapt before
measurements were recorded for three consecutive 24h days starting at midnight. On the morning of the last
3347
Table 1. Ingredient and chemical composition (DM basis)
of the diet used in Exp. 1 and 2
Item
Ingredienta
Barley silageb
Barley grain, steam-rolled
Barley grain, ground
Canola meal
Urea
Trace mineral and vitamin premixc
Diclacium phosphate
Calcium carbonate
Salt
Canola oil
Dried molasses
Flavoring agentd
Chemical compositione
DM, %
OM, % of DM
GE, Mcal/kg
CP, % of DM
NDF, % of DM
ADF, % of DM
Ca, % of DM
P, % of DM
%
75.0
19.0
1.03
1.52
0.59
0.067
0.30
0.76
1.52
0.06
0.15
0.003
44.3 ± 0.1
91.9 ± 0.4
4.65 ± 0.04
14.5 ± 0.5
36.2 ± 2.2
16.9 ± 1.2
0.50
0.23
a
All ingredients except for rolled barley and silage were provided
as part of a mash supplement.
b
Composition was 92.5 ± 0.8% OM, 11.9 ± 1.1% CP, 42.1 ± 1.6%
NDF, and 21.1 ± 0.9% ADF based on four samples composited by
period.
c
Supplied per kilogram of dietary DM: 65 mg/kg of Zn; 28 mg/kg
of Mn; 15 mg/kg of Cu; 0.30 mg/kg of Se; 0.2 mg/kg of Co; 0.7 mg/kg
of I, 6,000 IU of vitamin A; 600 IU of vitamin D; and 14 IU of vitamin
E.
d
Anife 420 Power, Canadian Biosystems, Inc., Calgary, Alberta,
Canada.
e
Mean and standard deviation for four pooled samples (each period),
except for minerals, which were analyzed using a pooled sample for
the experiment.
day of each period, the cattle were removed from the
chambers and transported to their individual stalls in
the barn. Interruptions to the chamber flux measurements occurred daily at 0730, when the floor was
cleaned, when fecal samples were taken at 0930, when
cattle were fed, and at 1530, when fecal samples were
again taken.
Cattle, Diet, and Treatments
Sixteen Holstein steers weighing 311.6 ± 12.3 kg at
the start of the experiment were used (eight steers per
experiment). The cattle were selected for their temperament and then conditioned to the metabolism stalls
over a period of several months before experimentation.
This was done to minimize the stress on the cattle during the experiments.
During the experiments, the steers received a basal
diet containing 75% whole crop barley silage, 19%
steam-rolled barley, and 6% supplement (DM basis).
Diet composition is shown in Table 1. The diet was
formulated using NRC (1996) recommendations to contain 14% CP, and to meet or exceed the mineral and
vitamin requirements of cattle gaining 1 kg/d.
3348
McGinn et al.
In Exp. 1, treatments were control (no additive), monensin (Rumensin, Elanco Animal Health; 33 mg/kg
DM), sunflower oil (400 g/d, approximately 5% added
fat), and proteolytic enzyme (Protex 6-L, Genencor Int.,
Inc.; 1 mL/kg DM). Sunflower oil is rich in oleic (45.3%)
and linoleic (39.8%) acid (NRC, 2001). The level of fat
added to the diet was close to the maximum range typically recommended for use in cattle diets (NRC, 2001).
The source organism for the enzyme product was Bacillus licheniformis and the protease activity was determined to be 4,507 U/mL (SD = 161.0; n = 5). Protease
activity was determined at pH 6.0 and 39°C (i.e., ruminal conditions) using 0.4% (wt/vol) azocasein as substrate (Bhat and Wood, 1989). Briefly, a reaction mixture containing 0.5 mL of azocasein, 0.5 mL of citratephosphate buffer, and 25 ␮L of enzyme (diluted 1:100
in distilled water) was incubated at 39°C for 15 min.
The unhydrolyzed azocasein was precipitated by adding
80 ␮L of 25% (wt/vol) trichloroacetic acid and then removed by centrifugation (2,040 × g for 10 min at room
temperature). A 0.5-mL supernatant sample was mixed
with 0.5 mL of 0.5 M NaOH and the absorbance read at
420 nm against a reagent blank. Enzyme (no substrate)
and substrate (no enzyme) blanks were also included
for correction. One unit of protease activity was defined
as the absorbance measured at 420 nm by the action
of 10 ␮g of a standard protease (Streptomyces griseus,
Type XIV, Sigma Chemicals, St. Louis, MO), assayed
under identical conditions. The enzyme product contained negligible cellulase (using carboxymethylcellulose), xylanase (using xylan from birchwood), and αamylase (using soluble starch) activities when measured under ruminal conditions (39°C and pH 6.0).
In Exp. 2, treatments were: control (no additive); Procreatin-7 yeast (SAF Agri; 4 g/d; Saccharomyces cerevisiae; 1.5 × 1010 cfu/g); Levucell SC yeast (Lallemand,
Inc.; 1 g/d; Saccharomycces cerevisiae, strain CNCM I1077; 2 × 1010 cfu units/g); and fumaric acid (Bartek
Ingredients Inc.; 80 g/d). The feeding rates used for the
yeast products were those recommended by the respective suppliers.
In both experiments, the treatments were hand
mixed into the diet at the time of feeding. The basal diet
was prepared daily using a feed mixer (Data Ranger,
American Calan, Inc., Northwood, NH) and feed was
offered once daily for ad libitum intake (at least 10%
orts). Quantities of feed offered and refused were recorded daily for each animal. Samples of diet and refusals were retained weekly for determination of DM content. The DMI was calculated daily for each steer as
the DM offered minus the DM refused.
Chamber Design
The calculation of CO2 and CH4 emissions was based
on their respective concentration measurements associated with airflows into and out of each chamber. The
chambers were 4.4 m wide × 3.7 m deep × 3.9 m tall (63.5
m3; model C1330, Conviron Inc., Winnipeg, Manitoba,
Figure 1. Ventilation design of cattle chamber used to
measure gas emission.
Canada). Within each chamber, the animals were individually restrained in metabolism stalls that measured
2.5 m long × 0.9 m wide, elevated from the floor by 15 cm.
The ventilation of each chamber consisted of individual fresh-air intakes and chamber exhaust ducts (i.d.
30.5 cm), with dedicated fans in each duct. The air
volume of each chamber was exchanged approximately
every 5 min. Fresh intake air (0.28 m3/s) was fed directly
into a sealed box containing a pair of fans (squirreltype) that recycled the air inside the chamber (Figure
1). The recycled air entered the chamber through three
raised floor vents running the length of the chamber,
located between the animal stalls and between the
stalls and walls of the chamber. The recycled air was
filtered before leaving the chamber through vents above
the stalls. Between the air filter located in the room
and the low-pressure side of the recycling fans was a
condenser unit that kept the recycled air at a constant
temperature (15°C). The well-mixed air inside the
chamber was essential to ensure a representative sample of air through the chamber exhaust duct (0.22 m3/s).
An automated actuator device on each chamber, which
normally allowed some fresh intake air to be diverted
to the exhaust, was closed so that all fresh air entered
the chamber.
Below and to the rear of each stall was a hole with
a graded floor leading to the manure removal track.
Between the chamber and the track was a flexible rubber mat hanging vertically, the function of which was
to seal the chamber from the manure removal system.
The seal on these flaps was tested by monitoring pressure and flow changes in each chamber before and after
resealing the pit on the chamber-side of the flap. In
addition, the inside of each chamber was set at a constant positive pressure (3 to 5 Pa; model PX653-0.1D5V,
Omega Engineering Inc., Stamford, CT) to ensure no
inflow of gases through leaks in the chamber.
Chamber Flow and Concentration Measurements
The air velocity (model 8330, TSI Inc., Shoreview,
MN) and air temperature (thermocouple junction with
shielded cable) were measured in the fresh-air intake
3349
Methane emissions from beef cattle
and chamber exhaust ducts of each chamber. The air
velocity measurements were made manually three
times daily during the time the cattle were housed in
the chambers. This was necessary to confirm the consistency of air exchange throughout the day. An average
of five points across the diameter of each duct was used
in characterizing air velocity.
The concentration (ppm, volume basis) of CO2 in the
intake and exhaust ducts of each chamber was measured by pumping a sample of the air stream in each
duct through infrared gas analyzers. Each chamber was
equipped with a dedicated CO2 gas analyzer. For the
first 5 min, the intake air stream was sampled for chambers 1 and 2 (using dedicated analyzers 1 and 2),
whereas the other two analyzers (3 and 4) sampled
the exhaust air stream of chambers 3 and 4. For the
subsequent 5 min, the system was reconfigured so that
analyzers 1 and 2 sampled the exhaust air stream,
while analyzers 3 and 4 sampled intake air streams.
In this manner, both the intake air and exhaust air
were sampled in each chamber every 10 min using the
same analyzer for each chamber. The CO2 concentration analog output from each analyzer was sampled
every 5 s and the average recorded every 5 min using
a datalogger (model CR23X, Campbell Scientific Inc.,
Logan, UT). Three of the CO2 analyzers (model LI-6262,
LI-COR Inc., Lincoln, NE) also measured dew point
temperature (Td) and were programmed to correct CO2
concentration for dilution and pressure broadening effects due to water vapor in the air stream. The fourth
CO2 analyzer (model LI-6252, LI-COR Inc.) was not
capable of monitoring H2O directly but was fed an analog Td signal from the companion analyzer. The CO2
concentration and Td measured by each analyzer required barometric pressure that was monitored with an
external pressure transducer (model CS105 barometer,
Vaisala Inc., Vantaa, Finland). The pressure transducer was wired to the datalogger that disseminated
the barometric value via an analog signal back to each
CO2 analyzer.
The concentrations (ppm) of CH4 for each chamber
were monitored differently than those of CO2. In this
case, only the fresh air intake concentration was sampled using an infrared gas analyzer (model Ultramat
5E, Siemens Inc., Karlsruhe, Germany) with a 0 to 50
ppm range. Air was pumped (1 LPM; model TD3LS7,
Brailsford and Co., Inc., Rye, NY) sequentially from
each of the chamber’s fresh-air intakes for 15 min, and
the logger recorded the average concentration every 5
min. The first 5 min of each 15-min sampling interval
was ignored because this corresponded to the time required by the analyzer to adjust to the new air stream
concentration. The CH4 concentration of the chamber
exhaust air was measured indirectly using an open path
laser (GasView MC, Boreal Lasers Inc., Spruce Grove,
Alberta, Canada) mounted inside each chamber between the animal stalls at 1.5 m above the central floor
return-air vent. Exhaust methane concentration, measured periodically with the infrared methane analyzer,
Table 2. Relationship between the open-path laser (chamber air) and infrared gas analyzer (exhaust duct) methane concentrations
Chamber
Slope
Intercept
Pearson
correlation
Concordance
correlationa
1
2
3
4
1.306
1.373
1.146
1.182
−1.6
−2.1
−0.7
−0.5
0.99
0.99
0.99
0.99
0.93
0.89
0.98
0.97
a
Measure of closeness to 1:1 line (Lin, 1992).
was regressed against chamber concentration, measured using the laser. A consistent relationship was
found between each chamber’s CH4 concentration (wellmixed air) and each chamber’s exhaust air CH4 concentration (Table 2). These data showed a high degree of
precision (Pearson correlation) and accuracy (Concordance correlation; Lin, 1995) between the chamber laser methane concentrations and that measured in the
exhaust duct with the infrared methane analyzer.
These corrections were applied to the chamber laser
concentration data to derive exhaust concentration.
Once each morning before the time the cattle were
fed, the gas analyzers were checked for calibration (zero
with N2 and span with a standard gas) causing a loss
of emission data for up to 30 min each day.
Flux Calculations
The CH4 emission generated for each chamber (FCH4;
g/s) was calculated for each 10-min period from the
fresh-air intake (i) and chamber exhaust (e) concentration (Ci and Ce, respectively; ppm) and mean weekly
air velocity (Vi and Ve, respectively; m/s) data:

T  
T 
P
P
FCH4 = Ce MW
Ve A − Ci MW
Vi A 
RT
P  
RT
P 

where MW is the molecular weight of CH4 (16 g/mol),
P is the barometric pressure (Pa), R is the universal
gas constant (8.31 Jⴢmol−1ⴢdeg K−1), T is the stream air
temperature (°K), and A is the cross sectional area of
the duct (0.146 m2). The T/P-value is a correction for
the air velocity meter. The calculation of CO2 emissions
(FCO2) for each chamber was determined using the same
equation, where MW was 14 g/mol.
Errors in the emission calculation were associated
with the measurement sensitivity of the gas concentration analyzers and the measurement sensitivity and
temporal variability in air velocity within the intake
and exhaust ducts. The open-path laser specified a sensitivity of ± 0.27 ppm, whereas that for the infrared gas
analyzer was ± 0.25 ppm. The error in air velocity meter
was ± 0.025 m/s, and the error in assuming constant
airflow over 3-d measurement periods was ± 0.14 m/s.
The probable error analysis of these individual terms
in the flux equation indicates an overall error in flux
3350
McGinn et al.
estimates of 7%; however, these errors were consistent
among chambers implying treatment sensitivity would
be much greater.
Correction factors accounting for between-chamber
differences were applied to all flux data, decreasing
the chamber effect to a random effect in the statistical
analysis. The correction factors were developed by releasing a controlled amount of pure CH4 from a gas
cylinder at the same rate into each empty chamber
(sequentially). The flux of CH4 from each chamber was
determined when the exhaust concentration reached
steady state. The test was conducted three times during
the experiment, and the ratio of maximal flux (always
chamber 4) to chamber flux was determined. The average correction factors relative to chamber 4 were 1.01,
1.14, 1.13, and 1.00 for chambers 1 to 4, respectively.
These chamber corrections were also applied to the CO2
flux data. This procedure decreased the variability in
emissions attributed to chambers to within 3%, and
improved the sensitivity to treatment differences.
Ruminal Fermentation Measurements
Ruminal pH was measured once per animal on d 14
of the period 3.5 h after feeding. A rubber tube was
inserted into the rumen via the esophagus and rumen
contents (400 mL) removed using an electric pump.
Samples were monitored visually to ensure they were
not contaminated with saliva. The pH was measured
immediately using a pH meter (Accumet model 25, Denver Instrument Co., Arvada, CO). The whole contents
were squeezed through four layers of cheesecloth. Five
milliliters of the filtrate was combined with 1 mL of
25% (wt/vol) of meta-phosphoric acid and stored frozen
(−30°C) until VFA analysis.
Digestibility
Total-tract digestibility of nutrients was determined
using an external marker. The marker was prepared
using chromic oxide and ground barley. Ten grams of
marker, providing approximately 2 g of Cr, was top
dressed once daily onto the feed the last 10 d of each
period. In all cases, the entire allotment of marker was
consumed. A representative sample of the marker was
retained each period for Cr analysis. Fecal samples (100
g wet weight) were collected twice daily during the last
4 d of each period. During this time, the cattle were
housed in metabolism stalls within the chambers; thus,
the samples were obtained from the rectum of each
animal or from the floor when rectal samples were not
available. Fecal samples were composited by steer and
period as collected, and immediately frozen. The pooled
samples were later dried at 55°C for 48 h in a forcedair oven, ground through a 1-mm screen, and analyzed
for analytical DM, GE, NDF, ADF, and Cr. An additional fecal sample (100 g wet weight) was also taken
from each animal before dosing the marker each period.
These samples were analyzed for DM and Cr. The Cr
concentration in the feces taken predosing was used to
adjust for residual marker excretion. The Cr concentration in fecal samples obtained before dosing was in all
cases less than 1% of the Cr concentration in the fecal
samples composited by period; thus, this adjustment
was trivial. Chromium was assumed to be completely
indigestible and the digestibility of DM was calculated
as follows:
DM digestibility = 1
− [Cr fed (mg/d)/ DMI (kg/d)]/Cr in feces (mg/kg DM);
where DMI was the DM consumed on the same days
that fecal samples were collected. Digestibility of GE,
NDF, and ADF was calculated using the same approach.
Chemical Analyses
All chemical analyses were performed on each sample
in duplicate, and where the coefficient of variation for
the replicate analysis was >5%, the analysis was repeated.
Ruminal VFA were quantified using colonic acid as
the internal standard, and gas chromatography (model
5890, Hewlett Packard, Little Falls, DE) with a capillary column (30 m × 0.25 mm i.d., 1 ␮m phase thickness,
bonded polyethylene glycol, Supelco Nukol, Sigma-Aldrich Canada, Oakville, Ontario, Canada), and flame
ionization detection. The oven temperature was 100°C
for 1 min, which was then ramped by 20°C/min to 140°C,
and then by 8°C/min to 200°C/min, and held at this
temperature for 5 min. The injector temperature was
200°C, the detector temperature was 250°C, and the
carrier gas was helium.
Analytical DM was determined by drying the samples
at 135°C for 2 h, followed by hot weighing. The OM
content was calculated as the difference between 100
and the percentage of ash (AOAC, 1995; Method 942).
Gross energy was determined using an adiabatic calorimeter (model 1241, Parr, Moline, IL). The NDF and
ADF were determined in the ANKOM200 fiber analyzer
(Ankom Technology Corp., Fairport, NY) using heat
stable α-amylase and sodium sulfite. For the measurement of CP (N × 6.25), samples were ground using a
ball mill (Mixer Mill MM2000; Retsch, Haan, Germany)
to a fine powder. Nitrogen was quantified by flash combustion with gas chromatography and thermal conductivity detection (Carlo Erba Instruments, Milan, Italy).
Chromium, Ca and P were determined by inductively
coupled plasma emission spectrometry (SpectoCirosCCD, Specto Analytical Instruments, GmbH & Co.,
Kleve, KG, Germany) after dry ashing and extraction
of the respective mineral.
Calculations and Statistical Analyses
Cumulative daily CH4 emissions from each chamber
were calculated for 3 d each period. The daily CH4 flux
3351
Methane emissions from beef cattle
Table 3. Ad libitum dry matter intake and ruminal fermentation variables for cattle fed
a high-forage diets with various supplemental additives (Exp. 1, n = 8)
Treatments
Item
Control
Enzyme
Monensin
Oil
SEM
Treatment
P-value
Ad libitum DMI, kg/d
Ruminal pH
Total VFA, mM
Branch-chain VFA, mM
VFA, %
Acetate
Propionate
Butyrate
Iso-butyrate
Valerate
Iso-valerate
Caproate
Acetate:propionate
8.73
7.04b
68.1
1.66a
9.16
6.80a
82.0
1.80ab
9.02
6.78a
79.0
2.28c
8.70
6.75a
77.7
2.20bc
0.40
0.11
6.1
0.22
0.11
0.04
0.25
0.02
68.0c
18.3a
9.9
0.92
1.03
1.48a
0.38bc
3.75b
66.3bc
19.7ab
10.2
0.90
1.14
1.27a
0.47c
3.40ab
65.5ab
21.6b
8.6
1.01
1.10
1.88b
0.26a
3.06a
64.0a
21.2b
10.4
1.01
1.16
1.90b
0.28ab
3.10a
0.9
0.9
0.6
0.07
0.06
0.16
0.06
0.17
0.01
0.03
0.09
0.13
0.10
0.005
0.001
0.02
Within a row, means without a common superscript letter differ, P < 0.05.
a,b,c
(13.3 Mcal/kg CH4) determined for each chamber was
expressed as a proportion of GE intake and DE intake
of the two cattle within the chamber on that same day.
The daily CH4 flux was also expressed per unit of DMI
for the two cattle within the chamber on that same day.
One animal fed the Levucell SC yeast in Period 3
went off-feed while in the chamber. Thus, for Period 3,
the gas emission data for this chamber were removed
from the analysis. Data were analyzed for each experiment with the mixed model procedure of SAS (SAS
Inst., Inc., Cary, NC). The individual animal was the
experimental unit for intake, digestibility, and ruminal
fermentation variables because these data were obtained from individual animals with separate access
to the feed. The chamber, representing data for two
animals, was the experimental unit for CH4 and CO2
measurements. The model for intake, digestibility, and
ruminal fermentation variables included the fixed effects of treatment. Animal and period were considered
as random effects and the restricted maximum likelihood method was used to estimate the variance components. The model used for CH4 measurements included
the fixed effect of treatment, and the random effects of
period and chamber, with day of sampling (1 to 3) within
each period treated as a repeated measure. Differences
among means were tested using a protected (P < 0.05)
LSD test. Treatment effects were declared significant
at P < 0.05 and trends were discussed at P < 0.15.
Results
Dry Matter Intake
While the steers were housed in individual pens outside the chambers, their ad libitum DMI averaged 8.88
kg/d for those in Exp. 1 (Table 3) and 8.89 kg/d for those
in Exp. 2 (Table 4). Overall, treatments had no effect
(P = 0.11) on DMI, although there was a trend for higher
intake when the animals were fed enzyme or monensin
compared with the control diet (Table 3). Moving the
steers to the chambers decreased intake by 15% in Exp.
1 (Table 5) and by 19% in Exp. 2 (Table 6), such that
intake averaged 7.54 and 7.16 kg/d, respectively. It is
speculated that the drop in intake when steers were
moved from individual pens to metabolic stalls in chambers was due to the stress associated with the change
in environment, as well as the decreased energy expenditure due to the decreased activity of the cattle during
the time they were in the metabolic stalls. Once in the
chambers, steers receiving Levucell SC yeast had lower
(P = 0.02) DMI than the control group, but the intake
of the cattle fed the other treatments was not different
(P > 0.05) from the control group (Table 6).
Ruminal Fermentation
Ruminal pH was lower (P = 0.04) for steers fed monensin, enzyme, and sunflower oil compared with steers
fed the control diet; however, these pH values were all
well within the range required to maximize ruminal
fiber digestion (Table 3). Although there were no differences in total VFA concentration (mean of 76.7 mM)
among treatments in Exp. 1, proportions of individual
VFA differed. Steers fed either monensin or sunflower
oil had higher (P = 0.03) concentrations of propionate
and lower (P = 0.01) concentrations of acetate than
those in the control group. Consequently, acetate:propionate ratios averaged 3.06 for the monensin treatment
and 3.10 for the oil treatment compared with 3.75 for
the control group (P < 0.05). Concentrations of acetate,
propionate, and acetate:propionate ratio were intermediate (P > 0.05) for the steers receiving enzyme. In
addition, branch chain VFA and iso-valerate concentrations were higher (P < 0.05), and caproate concentration
was lower (P = 0.001), for steers receiving monensin or
oil than for those in the control group. The yeast and
fumaric acid treatments used in Exp. 2 had no effect
on ruminal pH or VFA (Table 4).
3352
McGinn et al.
Table 4. Ad libitum dry matter intake and ruminal fermentation variables for cattle fed
a high-forage diets with various supplemental additives (Exp. 2, n = 8)
Treatmentsa
Item
Control
Fumaric
acid
Levucell SC
yeast
Procreatin-7
yeast
SEM
Treatment
P-value
8.84
6.92
73.6
1.75
8.66
6.86
74.5
1.80
8.89
6.97
72.4
1.66
9.14
6.99
69.1
1.69
0.46
0.13
6.4
0.19
0.58
0.69
0.77
0.77
67.9
18.9
9.3
0.93
1.18
1.42
0.43
3.63
66.8
18.8
10.3
0.93
1.20
1.47
0.42
3.59
67.1
19.3
9.7
0.91
1.16
1.36
0.45
3.52
66.8
19.4
9.7
0.94
1.20
1.53
0.42
3.50
1.2
0.9
0.6
0.04
0.08
0.09
0.06
0.22
0.30
0.75
0.25
0.89
0.96
0.11
0.87
0.83
Ad libitum DMI, kg/d
Ruminal pH
Total VFA, mM
Branch chain VFA, mM
VFA, %
Acetate
Propionate
Butyrate
Iso-butyrate
Valerate
Iso-valerate
Caproate
Acetate:propionate
No significant effect of treatments for any of the variables measured, P > 0.15.
a
Table 5. Nutrient intake and digestibility in the total tract measured in cattle housed in
chambers and fed a high-forage diet with various supplemental additives (Exp. 1, n = 8)a
Treatments
Item
Intake, kg/d
DM
GE
NDF
ADF
Digestibility, %
DM
GE
NDF
ADF
Control
Enzyme
Monensin
Oil
SEM
Treatment
P-value
7.46
34.65
2.69
1.26
7.60
35.31
2.73
1.27
7.75
36.03
2.81
1.31
7.35
36.14
2.50
1.16
0.43
1.90
0.13
0.06
0.56
0.67
0.06
0.06
62.0b
61.6bc
44.3bc
35.7bc
56.8c
56.6c
37.1cd
29.1c
63.8b
63.6b
47.8b
40.3b
58.2bc
57.9c
34.1d
25.4c
2.4
2.2
4.5
5.4
0.04
0.04
0.02
0.04
a
Nutrient intakes and digestibility determined for 4 d, during which the animals were in the chambers.
Within a row, means without a common superscript letter differ, P < 0.05.
b,c,d
Table 6. Nutrient intake and digestibility in the total tract measured in cattle housed in
chambers and fed a high-forage diet with various supplemental additives (Exp. 2, n = 8)a
Treatments
Item
Intake, kg/d
DM
GE
NDF
ADF
Digestibility, %
DM
GE
NDF
ADF
a
Control
Fumaric
acid
Levucell SC
yeast
Procreatin-7
yeast
SEM
Treatment
P-value
7.38cd
34.45cd
2.67b
1.24b
6.95bc
32.42bc
2.50c
1.16c
6.79b
31.67b
2.43c
1.13c
7.51d
35.07d
2.72b
1.27b
0.27
1.23
0.06
0.03
0.02
0.02
0.005
0.007
63.9
63.8
45.8
39.7
62.0
61.9
42.3
34.9
64.6
65.1
46.2
38.2
64.6
64.5
45.3
38.0
2.3
2.2
2.7
3.2
0.70
0.62
0.74
0.74
Nutrient intakes and digestibility determined for 4 d, during which the animals were in the chambers.
Within a row, means without a common superscript letter differ, P < 0.05.
b,c,d
3353
Methane emissions from beef cattle
Table 7. Daily methane emissions from cattle fed a high-forage diet with various supplemental additives (Exp. 1, n = 4)a
Treatments
Item
DMI, kg/d
Methane
g/steer
g/kg of DMI
% GE intake
% DE intake
Control
7.40
166.2b
22.64b
6.47b
10.51bc
Enzyme
Monensin
7.55
164.4b
22.11b
6.32b
11.27b
7.71
159.6b
20.70b,e
5.91b,e
9.31cd,e
Sunflower
oil
6.91
129.0c
18.81c
5.08c
8.76d
SEM
Treatment
P-value
0.45
0.36
8.0
1.16
0.31
0.56
0.004
0.02
0.008
0.01
a
Methane emissions and corresponding DMI determined for 4 d, during which the animals were in the
chambers.
b,c,d
Within a row, means without a common superscript letter differ, P < 0.05.
e
The mean for monensin tended to differ from the control, P < 0.10.
Digestibility
Supplementing the diet with sunflower oil increased
the intake of GE by 1.5 Mcal/d (P = 0.29) compared
with the control, but this difference was not statistically
significant (Table 5). However, intakes of NDF and ADF
tended (P = 0.06) to be lower for animals fed oil because
of the dilution effect of adding 5% oil to the diet.
Adding oil to the diet had a negative effect on fiber
digestion. Compared with the Control, the NDF digestibility was decreased by 23% (P = 0.03), and ADF digestibility was reduced by 29% (P = 0.06) as the result of
feeding oil. Digestibilities of DM (P = 0.14) and GE (P =
0.15) for cattle fed oil were numerically lower but not
different than for the control.
Supplementing the diet with proteolytic enzyme decreased (P = 0.05) digestibility of DM in the total tract
by 8%. This was mostly the result of a reduction in
fiber digestibility. The NDF digestibility of cattle fed
enzymes was 14% lower (P = 0.10) and ADF digestibility
was 15% lower (P = 0.22) than for control animals,
although these differences were not statistically significant.
Monensin had no effect on tract digestibility of nutrients. Furthermore, the yeast and fumaric acid treatments used in Exp. 2, had no effect on digestibility of
nutrients in the total tract (Table 6).
Methane Emissions
Before correcting CH4 emissions for DMI or energy
intake (Table 7), sunflower oil reduced (P = 0.001) CH4
by 22% compared with the control (Exp. 1). In contrast,
monensin (P = 0.44) and enzyme (P = 0.82) had no effect
on CH4 emissions. When CH4 emissions were corrected
for differences in feed intake, the effect of sunflower oil
on lowering (P = 0.006) CH4 emissions was maintained.
Monensin also tended (P = 0.08) to lower CH4 emissions
per kilogram of DMI by 8.6% compared with the control
due to numerically higher intake and numerically lower
CH4 emissions.
For the high-forage backgrounding diet used in this
study, approximately 6.5% of the GE consumed, and
10.5% of the DE consumed, was lost in the form of CH4
emissions. Supplementing the diet with sunflower oil
decreased the loss of dietary energy by 21% in the case
of GE intake (P = 0.002) and 17% in the case of DE (P =
0.02). Monsensin tended to reduce the loss of dietary
energy by 9% in the case of GE intake (P = 0.09) and
11% in the case of DE (P = 0.09).
The yeast and fumaric acid treatments used in Exp.
2 had no effect (P > 0.05) on CH4 emissions (Table 8).
The CH4 emissions per kilogram of DMI (P = 0.39)
and as a percentage of GE intake (P = 0.39) were not
significantly different for steers fed Procreatin-7 yeast
compared with control steers (3% less CH4 per kilogram
of DMI and 3% less GE intake was lost as CH4).
Carbon Dioxide Emissions
There were no differences (P > 0.05) due to treatment
in the amount of CO2 respired daily in Exp. 1 (average =
3.11 kg/animal; data not shown). However, in Exp. 2,
the fumaric acid and Levucell SC yeast increased daily
CO2 (3.97 and 4.02 kg/animal, respectively) respiration
(P < 0.05) relative to the control (3.61 kg per animal;
data not shown).
Discussion
The basal diet used in this study was typical of diets
fed in western Canada to backgrounded cattle. This
forage-based diet contained high-quality barley silage
and barley grain and was estimated to sustain a growth
rate of about 1 kg/d (NRC, 1996). Actual growth rate
during the study averaged 0.92 ± 0.10 kg/d for steers
in Exp. 1 and 0.83 ± 0.15 kg/d for those in Exp. 1. Effects
of treatments on gain were not reported as they were
thought not to be representative of the commercial situation because the animals were restrained while in
the chambers.
For cattle fed the control diet, about 6.5% of the ingested energy was lost as CH4. These data confirm the
report by Johnson and Johnson (1995) that feedlot cattle fed backgrounding-type diets and replacement cattle
3354
McGinn et al.
Table 8. Daily methane emissions from cattle fed a high-forage diet with various supplemental additives (Exp. 2, n = 4)a
Treatments
Item
DMI, kg/d
Methane
g/steer
g/kg of DMI
% of GE intake
% of DE intake
Control
Fumaric
acid
Levucell SC
yeast
Procreatin-7
yeast
SEM
Treatment
P-valueb
7.18
6.69
6.71
7.46
0.54
0.06
179.0
25.05
7.13
11.36
171.7
26.00
7.40
12.02
171.9
26.43
7.53
11.81
180.9
24.32
6.93
10.77
9.9
1.78
0.50
1.10
0.30
0.15
0.15
0.28
a
Methane emissions and corresponding DMI determined for 3 d during which the animals were in the
chambers.
b
No significant effect of treatments for any of the variables measured, P < 0.05.
fed high forage diets typically lose 6 to 6.5% of their
ingested energy as CH4.
Adding sunflower oil to the diet substantially decreased CH4 emissions, corroborating numerous previous studies that reported reduced CH4 emissions using
other types of fats (Machmüller and Kreuzer, 1999;
Dohme et al., 2000). Adding fat to ruminant diets has
been shown to decrease CH4 losses mainly by decreasing ruminally fermentable substrate, but also by providing an alternative H sink in the rumen and by inhibiting protozoa (Johnson and Johnson, 1995). In our
study, adding sunflower oil to the diet clearly decreased
ruminal fermentability of the fiber as evidenced by
lower acetate concentration, higher propionate concentration, and a lower acetate:propionate ratio. Totaltract digestibility of fiber was also substantially reduced, even though decreases in total tract fiber digestion are usually not as severe as reductions in ruminal
fiber digestion. Thus, although ruminal fiber digestion
was not measured directly in this study, other variables
indicated considerable reduction in ruminal fiber digestion. The reduction in ruminal fiber digestion likely
accounted for a large proportion of the reduction in CH4
emissions. Decreased fiber digestion due to added fat
was reported previously for high levels of added fat
(Jenkins, 1993). The mechanism for reduced fiber digestibility due to added fat may be related to the process
of hydrogenation of the unsaturated fatty acids in the
rumen. If the ability of the microorganisms to saturate
the fatty acids is exceeded, then unsaturated fatty acids
accumulate and interfere with microbial digestion
(NRC, 2001). In our study, the DE content of the diet
was not increased by the addition of 5% fat because the
additional DE supplied by the oil was offset by the
decrease in fiber digestion. Based on our results, adding
sunflower oil to the diet can be used to decrease CH4
emissions, but total energy intake may not be increased
because of negative effects on fiber use.
Monensin also tended to lower CH4 emissions, albeit
to a lesser extent than sunflower oil. Ionophores such
as monensin are typically used in commercial feedlot
cattle diets to modulate intake, control bloat, and improve feed efficiency (Elanco Animal Health, 2003). In
this study, monensin did not lower feed intake, but the
decreased concentration of acetate, increased concentrate of propionate, and reduced actetate:propionate
ratio were consistent with the known mode of action
of monensin (Schelling, 1984). The approximately 9%
decrease in CH4 emissions as a proportion of GE observed for monensin in this study is within the range
(slight to 25%) reported previously (Johnson and Johnson, 1995). Several recent studies have reported that
the effects of monensin on CH4 emissions are shortlived, as reviewed by Johnson and Johnson (1995). Our
results confirm that monensin in a high-forage diet is
a viable strategy to decrease CH4 emissions in the short
term; however, a feeding trial of greater length is required to determine the long-term effect of monensin
on CH4 emissions.
No other treatment significantly affected CH4 emissions, although the 3% decrease in feed energy lost
as CH4 observed for cattle fed Procreatin-7 yeast was
noteworthy. Although this reduction in CH4 was small
and not statistically significant, the cost of feeding yeast
was considerably less than the cost of feeding sunflower
oil (i.e., approximately one-seventh). In this case, the
decrease in energy lost as CH4 was the result of slightly
higher feed intake combined with similar CH4 emissions compared with the control group. The means by
which this yeast product might have decreased the proportion of feed energy lost as CH4 is not certain, but
was likely due to a shift in specific microbial populations
within the overall community in the rumen (Newbold
et al., 1996). There is some limited work to suggest that
live yeast cells can stimulate the use of H by acetogenic
strains of ruminal bacteria, thereby enhancing the formation of acetate and decreasing the formation of CH4
(Chaucheyras et al., 1995). However, in our study, there
were no differences in VFA concentrations, suggesting
that any possible changes in microbial fermentation
were too subtle to elicit a change in VFA concentrations.
The different response observed for the two yeast
products used in this study is not surprising. It is well
known that the effects of yeast are strain-dependent
(Newbold et al., 1996). The two products used in our
study differed in strain of Saccharomyces cerevisiae
3355
Methane emissions from beef cattle
used, as well as the number of yeast cells supplemented.
Although the Procreatin-7 yeast had no effect on ruminal fermentation or digestibility, it coincided with
slightly higher feed intake of cattle inside the chambers
during CH4. This finding lends support to the anecdotal
evidence observed commercially that yeast can have a
positive influence on feed intake and is useful in decreasing the effects of stress on feed intake in unadapted cattle.
The enzyme product used in this study was not effective in reducing CH4 losses. Colombatto et al. (2003a)
reported that in continuous culture, supplementing a
diet based on alfalfa hay, corn silage, and rolled corn
with a proteolytic enzyme enhanced ruminal fiber digestion without increasing CH4 emissions. Those findings contrast with the observations in the current
study, in which proteolytic enzyme had no effect on
total-tract fiber digestibility. The differences between
studies are attributed to the differences in diets used.
The key enzyme activities required to increase fiber
digestion depend on the composition of the diet on which
the enzymes are expected to act (Colombatto et al.,
2003b). Thus, a particular enzyme formulation is not
likely to be effective for all diets. Although protease
enzyme was found to be useful in enhancing degradability of alfalfa fiber (Colombatto et al., 2003a), the results
from this study suggest this is not the case for barley
silage. Similarly, Colombatto et al. (2003b) reported
that the enzyme products that increased degradation
of alfalfa hay were not the same ones that increased
degradation of corn silage.
Fumarate is a direct metabolic precursor of propionate, and thus, it has the potential to decrease methane
emissions by directing H into succinate rather than into
methane (López et al., 1999). However, at the level of
supplementation used in our study, fumaric acid was
not effective in reducing CH4 losses. This finding contrasts to in vitro (Asanuma et al., 1999; López et al.
1999) and in vivo (Bayaru et al. 2001) studies that
reported fumaric acid decreased CH4. It is possible that
a higher level of supplementation than that used in the
current study would be needed to alter CH4 production
in vivo. It is estimated that the level of supplementation
in this study provided about 15 mM of fumaric acid
(assuming a ruminal volume of 40 L; 116.07 g/mol),
although these calculations do not account for fluid dilution rate and passage of fumaric acid from the rumen
in vivo, factors that do not occur in vitro. In vitro, López
et al. (1999) used up to 10 mM, and Asanuma et al.
(1999) used up to 30 mM. In vivo, Bayaru et al. (2001)
fed about twice the amount of fumaric acid (20 g/kg
DMI) to cattle as that used in the current study (12 g/
kg DMI) and, surprisingly, methane production decreased by 23%. However, in that study, only two animals per treatment were used, and CH4 measurements
were made using ventilated hoods, which are known to
be problematic.
The lack of an effect of fumaric acid on CH4 emissions
was consistent with the lack of effect on VFA, notably
propionate proportions. The lack of an effect of fumaric
acid on ruminal pH was expected because of the highforage diet used. Previous studies have shown that organic acids can be beneficial in promoting higher ruminal pH and lactate uptake by ruminal organisms when
high-grain diets are fed (Martin et al., 1999). However,
with high-forage diets, lactate production in the rumen
is minimal; thus, organic acids are not expected to affect
ruminal pH.
The daily CO2 emission averaged 3.44 kg per animal
in our experiments, similar to the average CH4 emission
of 3.77 kg per animal, expressed as CO2 equivalent
(using a warming potential of 23 for CH4 relative to
CO2). Our CO2 emissions were similar to the 3.2 kg of
CO2ⴢanimal−1ⴢd−1 reported by Kinsman et al. (1995) for
dairy cows housed in a closed ventilated barn equipped
with instruments to directly measure CO2. However,
our CO2 emissions are greater than the 1.0 and 1.3 kg
of CO2ⴢanimal−1ⴢd−1 (using a conversion factor of 1,870
L/kg) reported for yearling beef heifers by Boadi et al.
(2002). The discrepancy between CO2 emissions reported in our study and those reported by Boadi et al.
(2002) is probably a reflection of the techniques used.
Boadi et al. (2002) used open-circuit calorimetry, which
consisted of a ventilated hood that enclosed the animal’s
head, as well as the sulfur hexafluoride tracer gas technique. Estimates of CO2 emissions were 20% lower using the hoods compared with the tracer technique, but
both techniques were considered highly variable. Despite CO2 emissions being large, the respired CO2 by
livestock is not reported by the Intergovermental Panel
on Climate Change (2001), presumably because it is a
rerelease of CO2 recently captured by photosynthesis.
Over our two experiments, an average loss of carbon
as CO2 and CH4 amounted to 1.97 kg of Cⴢanimal−1ⴢd−1.
In summary, approximately 6.5% of the GE consumed by growing cattle fed a high-forage, backgrounding diet is lost as CH4. Several feed additives and
ingredients that are currently registered for feeding to
cattle can be used to reduce the proportion of GE lost
as CH4. Methane emissions, expressed as a percentage
of GE, were decreased by 21% using sunflower oil and
by 9% using Rumensin. Additionally, one of the yeast
products examined numerically decreased CH4, as a
percentage of GE, by 3%. However, it must be acknowledged that this study was short-term, and that typical
production scenarios would not entail 3-wk feeding periods. It has been documented previously that that the
effectiveness of treatments, such as ionophores, in lowering CH4 can diminish over time (Johnson and Johnson, 1995). However, the objective of this study was
to identify additives and feed ingredients that were
effective, at least in the short-term, in decreasing CH4
emissions. The long-term effect of compounds identified
in this study will be examined in subsequent research.
Implications
This study demonstrates that several ingredients or
feed additives currently registered for feeding to beef
3356
McGinn et al.
cattle in Canada can be used to reduce the loss of energy
as methane. Specifically, sunflower oil and monensin
decreased methane emissions from cattle fed a highforage diet, although studies need to be done to confirm
whether such reductions are maintained over longer
feeding periods. Mitigating methane losses from cattle
will have long-term environmental benefits in terms
of reducing the contribution by animal agriculture to
greenhouse gas emissions; however, these ingredients
are expected to increase the cost of feeding. Whether
these ingredients are fed to commercial beef cattle will
depend on the economic benefits determined by their
effects on feed efficiency, rate of gain, and carcass characteristics.
Literature Cited
AOAC. 1995. Official Methods of Analysis. 16th ed. Assoc. Offic. Anal.
Chem., Arlington, VA.
Asanuma, N., M. Iwamoto, and T. Hino. 1999. Effect of the addition
of fumarate on methane production by ruminal microorganism
in vitro. J. Dairy Sci. 82:780–787.
Bayaru, E., S. Kanda, T. Kamada, H. Itabashi, S. Andoh, T. Nishida,
M. Ishida, T. Itoh, K. Nagara, and Y. Isobe. 2001. Effect of
fumaric acid on methane production, rumen fermentation, and
digestibility of cattle fed roughage alone. Anim. Sci. J.
72:139–146.
Benchaar, C., C. Pomar, and J. Chiquette. 2001. Evaluation of dietary
strategies to reduce methane production in ruminants: A modeling approach. Can. J. Anim. Sci. 81:563–574.
Bhat, M. K., and T. M. Wood. 1989. Multiple forms of endo-1,4-β-Dglucanase in the extracellular cellulase of Penicillium pinophilum. Biotechnol. Bioeng. 33:1242–1248.
Boadi, D. A., K. M. Wittenberg, and A. D. Kennedy. 2002. Validation
of the sulphur hexafluoride (SF6) tracer gas technique for measurement of methane and carbon dioxide production by cattle.
Can. J. Anim. Sci. 82:125–131.
Canadian Council on Animal Care. 1993. Guide to the Care and Use
of Experimental Animals. Vol. 1. 2nd ed. E.D. Olfert, B. M. Cross,
and A. A. McWilliam, ed. CCAC, Ottawa, ON, Canada.
Chaucheyras, F., G. Fonty, G. Bertin, and P. Gouet. 1995. In vitro
utilization by a ruminal acetogenic bacterium cultivated alone
or in association with an archaea methanogen is stimulated by
a probiotic strain of Saccharomyces cerevisiae. Appl. Environ.
Microbiol. 61:3466–3467.
Colombatto, D., G. Hervás, W. Z. Yang, and K. A. Beauchemin. 2003a.
Effects of enzyme supplementation of a total mixed ration on
microbial fermentation in continuous culture, maintained at
high and low pH. J. Anim. Sci. 81:2617–2627.
Colombatto, D., D. P. Morgavi, A. F. Furtado, and K. A. Beauchemin.
2003b. Screening of exogenous enzymes for ruminant diets: Relationship between biochemical characteristics and in vitro ruminal degradation. J. Anim. Sci. 81:2628–2638.
Dohme, F., A. Machmüller, A. Wasserfallen, and M. Kreuzer. 2000.
Comparative efficiency of fats rich in medium chain fatty acids to
suppress ruminal methanogenesis as measured with RUSITEC.
Can. J. Anim. Sci. 80:473–482.
Elanco Animal Health. 2003. http://www.elanco.com accessed December 16th, 2003.
Environment Canada. 2002. Canada’s Greenhouse Gas Inventory
1990–2000. Government of Canada Publication EPS 5/AP/8, Ottawa, ON, Canada.
Garcia-López, P. M., L. Kung, Jr., and J. M. Odom. 1996. In vitro
inhibition of microbial methane production by 9,10-anthraquinone. J. Anim. Sci. 74:2276–2284.
Jenkins, T. C. 1993. Lipid metabolism in the rumen. J. Dairy Sci.
76:3851–3863.
Johnson, K. A., and D. E. Johnson. 1995. Methane emissions from
cattle. J. Anim. Sci. 73:2483–2492.
Intergovernmental Panel on Climate Change. 2001. Page 881 in Climate Change 2001: The Scientific Basis. Cambridge Univ. Press,
Cambridge, U.K.
Itabashi, H. 2002. Reducing ruminal methane production by chemical
and biological manipulation. Page 139 in Greenhouse Gases and
Animal Agriculture. J. Takahashi and B.A. Young, ed., Elsevier
Sciences B.V., Amsterdam, The Netherlands.
Kinsman, R., F. D. Sauer, H. A. Jackson, and M. S. Wolynez. 1995.
Methane and carbon dioxide emissions from dairy cows in full
lactation monitored over a six-month period. J. Dairy Sci.
78:2760–2766.
Lin, L. I.-K. 1992. Assay validation using the concordance correlation
coefficient. Biometrics 48:599–604.
López, S., C. Valdés, C. J. Newbold, and R. J. Wallace. 1999. Influence
of sodium fumarate addition on rumen fermentation in vitro.
Br. J. Nutr. 81:59–64.
Machmüller, A., and M. Kreuzer. 1999. Methane suppression by coconut oil and associated effects on nutrient and energy balance in
sheep. Can. J. Anim. Sci. 79:65–72.
Martin, S. A., M. N. Streeter, D. J. Nisbet, G. M. Hill, and S. E.
Williams. 1999. Effects of DL-malate on ruminal metabolism and
performance of cattle fed a high-concentrate diet. J. Anim. Sci.
77:1008–1015.
Miller, T. L., and M. J. Wolin. 2001. Inhibition of growth of methaneproducing bacteria of the ruminant forestomach by hydroxymethylglutaryl-SCoA reductase inhibitors. J. Dairy Sci.
84:1445–1448.
Mutsvangwa, T., I. E. Edwards, J. H. Topps, and G. F. M. Paterson.
1992. The effect of dietary inclusion of yeast culture (Yea-Sacc)
on patterns of rumen fermentation, food intake and growth of
intensively fed bulls. Anim. Prod. 55:35–40.
Newbold, C. J., R. J. Wallace, and F. M. McIntosh. 1996. Mode of
action of the yeast Saccharomyces cerevisiae as a feed additive
for ruminants. Br. J. Nutr. 76:249–261.
NRC. 1996. Nutrient Requirements of Beef Cattle. 7th rev. ed. Natl.
Acad. Press, Washington, DC.
NRC. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl.
Acad. Sci., Washington, DC.
Schelling, G. T. 1984. Monensin mode of action in the rumen. J. Anim.
Sci. 58:1518–1527.