9

3
Research Article
Received: 1 December 2011 Revised: 5 April 2012 Accepted: 16 April 2012 Published online in Wiley Online Library: 1 June 2012
(wileyonlinelibrary.com) DOI 10.1002/jsfa.5735
Chemical composition and sensory quality of
bovine milk as affected by type of forage and
proportion of concentrate in the feed ration
Mette K Larsen,
a
Ulla Kidmose,
b
Troels Kristensen,
c
Pierre Beaumont
a
and Grith Mortensen
a
Abstract
BACKGROUND: The objective of this study was to investigate how some small changes in the forage content of maize and
lucerne silage and in the ration between forage and concentrate in the diet of dairy cows affect milk quality. Milk quality
was assessed by quantitative descriptive sensory analysis and by analysis of tocopherols and carotenoids as well as fatty acid
composition.
RESULTS: Changingthe ratiobetweenmaize silage andlucerne silage from5 : 1 to2 : 1 increasedmilk fat content of carotenoids
(2327%) and C18 : 3 n3 (15%), and reduced stale aroma and creamy avour. Increasing the proportion of concentrates in the
feed ration from 0.2 to 0.4 increased energy corrected milk yield (26%), reduced fat content (10%), increased C18 fatty acids
(862%) and reduced C16 (20%) content in milk fat. In addition, this milk type was described by the sensory panel as less oily,
less saturated and less yellow. The changes in milk composition were related to differences in feed composition.
CONCLUSION: The study revealed the potential to produce milk with a distinct composition and sensory quality based on even
small changes in the feed composition that are straightforward to implement by farmers.
c 2012 Society of Chemical Industry
Keywords: bovine milk; fatty acids; sensory evaluation; antioxidants; feeding
INTRODUCTION
Certaintypes of forage plants or other feedcomponents may inu-
ence milk composition and affect sensory as well as the nutritional
and technological quality of the dairy products. Additionally such
production conditions may affect the productivity and environ-
mental load. This may be used in the marketing of the produced
dairy products.
Numerous studies have been conducted to assess the effect
of various forage types on the milk composition, especially milk
fatty acid composition.
15
High contents of polyunsaturated fatty
acids, in particular C18 : 3 n3 and conjugated linoleic acid (CLA)
cis9 trans11 in milk fat, are desirable due to positive effects on
human health.
1,6
The main sources of polyunsaturated fatty acids
in normal cow diets are forage, especially grass crops (including
legumes such as clover and lucerne), and oil seeds or oil seed
by-products used as concentrates.
1,2,79
Only minor amounts of
the ingested polyunsaturated fatty acids from feed are recovered
in the milk as these fatty acids are hydrogenated to a wide extent
during the ruminal processes.
10
A higher content of polyunsaturated fatty acids makes the
milk more susceptible to oxidation, but this oxidation is delayed
to some extent by the inherent content of antioxidants such as
tocopherols, carotenoids and urate.
5,1114
Forages as well as oil seeds contain various amounts of
tocopherols and the main source of carotenoids is forage,
especially fresh grass and legumes and silage of grass and
legumes.
15
Urate is synthesised during the ruminal processes,
and the amounts are affected by the feed structure.
13
The sensory quality of milk is affected by direct transfer of
aroma compounds from feed to milk, but also by formation of
aroma compounds during feed digestion.
16
The effects of pasture
the composition of milk aroma compounds as well as the sensory
properties have been studied, and milk from cows consuming
pasture has a higher content of a range of aroma compounds,
whereas the sensory properties have been described as more
barny, or more salty, grassy, mothball and less sweet, sweet malty,
sweet aromatic compared to milk from cows fed a total mixed
ration(TMR).
1618
Compositionof theTMRalsoaffects milksensory
properties andmaizesilagecomparedtograss silagegives ahigher

Correspondence to: Mette K Larsen, Department of Food Science, Aarhus

University, DK-8830 Tjele, Denmark. E-mail: mette.larsen@agrsci.dk
Present address: VetAgroSup Clermont-Ferrand, Dept of Food Science, Site de
Marmilhat, FR-63370 Lempdes, France.
a Aarhus University, Department of Food Science, Blichers All e 20, DK-8830 Tjele,
Denmark
b Aarhus University, Department of Food Science, Kirstinebjergvej 10, DK-5792
Aarslev, Denmark
c Aarhus University, Department of Agroecology, Blichers All e 20, DK-8830 Tjele,
Denmark
J Sci Food Agric 2013; 93: 9399 www.soci.org c 2012 Society of Chemical Industry
9
4
www.soci.org MK Larsen et al.
sweet corn odor and less boiled milk, sour/buttermilk, cardboard
and metallic avour.
3,19
Feeding different lipid sources has only
minor effects on milk sensory properties.
5,19,20
These results indicate that the sensory quality of milk is highly
affected by forage composition, i.e. the type of forage as well as
the proportion of forage in the ration. Thus, the hypothesis of the
present work was that milk composition and sensory quality are
affected by even small variations in the proportion of forage and
concentrate as well as the variations in the composition of forage
and concentrates which could be observed between commercial
dairy farms.
EXPERIMENTAL
Animals, feeding and production
The experiment involved 72 Danish Holstein cows (mean at start:
86 37 days post partum, 30.1 5.7 kg energy corrected milk
yield) in a traditional production trial with eight treatments in
a balanced 2 2 2 design. Cows were randomly allocated to
treatments (nine cows per treatment) during an experimental
period of four weeks. The treatments were:
Ratio between maize silage and lucerne silage (5 : 1 or 2 : 1,
respectively)
Type of concentrate (barley or beet pulp in combination with
rape seed cake)
Proportion of concentrate (200 or 400 g kg
1
of planned dry
matter intake)
Maize was cut when the dry matter content in the total plant
reached 320 g kg
1
and ensiled in a bulk silo. Lucerne silage was
from the primary growth and cut when one tenth of the plants
were owering. After 1 day of wilting in the eld plants were
chopped and ensiled in a bulk silo.
All cows were housed in one group with ad libitum access
to one of the four different TMRs (Table 1) through Insentec
feeders (Insentec B.V., Marknesse, The Netherlands) controlling
the individual access and monitoring the intake. In addition, the
cows assigned to the high proportion of concentrate treatments
(400 g kg
1
of dry matter intake) were fed 2.2 kg dry matter
rapeseed cake in combination with either 2.2 kg dry matter barley
or 2.2 kg dry matter dried beet pulp at automatic feeding stations
in the barn controlling the individual access and the amount of
feed, the type of concentrate being identical to the type given in
the TMR for the individual cow.
Table 1. Composition (g kg
1
of dry matter) of the total mixed ration
fed ad libitum
Maizelucerne ratio 5 : 1 Maizelucerne ratio 2 : 1
Component Barley Beet pulp Barley Beet pulp
Barley 100 100
Beet pulp 100 100
Rapeseed cake 94 94 94 94
Maize silage
a
655 655 527 527
Lucerne silage
b
135 135 263 263
Urea 6 6 6 6
Mineral 10 10 10 10
a
DM 360 g kg
1
, ash 33, protein 85, starch 320, NDF 390 g kg
1
DM.
b
DM 300 g kg
1
, ash 128, protein 220, NDF 440 g kg
1
DM.
The composition of each TMR fed through the Insentec feeders
was measured daily. Individual intake of TMR as well as additional
concentrates was registered at each feeding bout and daily intake
was calculated. Individual milk recordings were performed six
times during the experimental period representing 24-h milk yield
froman afternoon andthe following morning milking. Milk fat and
protein analysis for registration of production was based on these
samples.
Sampling
Representative samples of each feed component were taken
weekly and were stored at 20

C until analyses of the fatty acid

composition and the content of tocopherols and carotenoids was
performed.
Individual milk samples were collected from two morning
milkings during the last week of the experiment and sub-samples
were frozen immediately after samplingandstoredat 20

Cuntil
analyses were performed for contents of tocopherols, uric acid
and carotenoids, fatty acid composition and colour measurement.
For sensory analysis, cows were milked in groups according to
treatments, and a 30 L milk sample was withdrawn from the bulk
milk of each treatment. Milk samples were pasteurised(72

C, 15 s)
within 6 h after milking, poured into 2-L milk plastic bottles that
were placed in ice water and kept at 1

C until sensory analysis

was carried out the next day.
Chemical analysis
Milk fat and protein contents were analysed using a Milkoscan
4000 (Foss, Hillerd, Denmark). The milk colour was measured
using a Minolta Chroma Meter (Minolta Co. Ltd, Osaka, Japan)
and the CIE (Commision Internationale dEclairage) Lab scale
with standardised daylight (D65). The L

, a

and b

values
reect lightness (0 = black, 100 = white), redness (100
= green, 100 = red) and yellowness (100 = blue, 100 =
yellow), respectively. The instrument was calibrated against a
white standard plate. Milk (14 mL) was poured into a white
porcelain basin (diameter 45 mm, height 30 mm) and colour was
measured on the liquid surface, i.e. where the front plate of
the instrument just touched the surface. Urate was analysed as
described previously.
13
Feed samples were mixed with liquid nitrogen and ground by
use of a domestic coffee grinder to obtain sizes of 2 mm length.
Carotenoids and tocopherols were extracted and analysed using
either 1 g ground feed sample or 2 mL milk sample as described
by Slots et al.,
21
with the following modication: HPLC analysis of
carotenoids was carriedout usinga YMCC30 column(YMCEurope
GmbH, Dinslaken, Germany), 250 4.6 mmi.d., 5 mparticle size,
operatedat atemperatureof 25

C. Isocraticelutionwas carriedout
with ethanol/methanol/tetrahydrofuran (75 : 20 : 5) (v/v/v) with a
ow rate of 1 mL min
1
, and chromatograms were recorded at
450 nm.
For fatty acids analysis, cream was separated from skim milk by
centrifugation (1700 g, 4

C, 20 min), subsequently cream was

centrifuged (13 000 g, 20

C, 10 min) to separate fat. A liquid

fat fraction was obtained by heating at 60

C for 10 min and

centrifugation (13 000 g, 40

C, 10 min). Ten milligrams of milk

fat was dissolved in 1 mL pentane and 10 L 12.5 mol L
1
sodium
methylate in methanol was added for methylation. Samples were
mixed for 1 min and left for 10 min at ambient temperature before
centrifugation (13 000 g, 4

C, 5 min). The supernatant was used

for analysis bygas chromatography. Fattyacids fromherbagewere
isolated and methylated as described by Palmquist and Jenkins.
22
wileyonlinelibrary.com/jsfa c 2012 Society of Chemical Industry J Sci Food Agric 2013; 93: 9399
9
5
Milk quality affected by forage and concentrate www.soci.org
Table 2. Feed intake and effect of ratio between maize and lucerne and share of concentrate on milk production
Proportion of silage,
maizelucerne
Amount of concentrate in
dry matter intake
Effect of silage Effect of amount
Parameter 5 : 1 2 : 1 proportion 20% 40% of concentrate
Feed intake (kg DM day
1
) 20.2 19.7 NS 18.1 21.7 <0.001
Total mixed ration
a
18.1 17.6 NS 18.1 17.6 NS
Rapeseed cake
b
1.1 1.1 NS 0 2.2 <0.001
Concentrate
b
1.0 1.0 NS 0 2.1 <0.001
Milk yield, energy corrected (kg day
1
) 28.0 28.5 NS 25.0 31.5 <0.001
Fat (g kg
1
) 38.8 42.0 <0.05 42.5 38.3 <0.01
Protein (g kg
1
) 30.7 30.4 NS 29.4 31.7 <0.001
a
Fed ad libitum;
b
Fed restricted.
NS, not signicant (P > 0.05).
Analysis of fatty acid methyl esters was carried out on an HP
6890 gas chromatograph tted with a ame ionisation detector
and a Restek RT 2560 capillary column (100 m, 0.25 m I.D.,
0.20 m df.).
21
Fatty acid methyl esters in milk were identied
using external standards (Supelco 37 component FAME mix from
Supelco, Bellafonte, USA; and PA and GLC 469 methyl ester
standard from Nu-Chek Prep Inc., Elysian, MN, USA) and amounts
calculated as milligrams per gram of identied fatty acids. Fatty
acids in herbage were identied by using external standards and
werequantiedby applyingaC17triglycerideas internal standard.
Sensory analysis
Atrainedsensory panel of 10 assessors evaluatedthe milk samples
in three sensory replicates. The temperatures of the milk samples
were 1618

C at serving. The milk samples were evaluated

using quantitative descriptive analysis. A sensory prole of 15
attributes for milk was assessed as follows: four aroma attributes,
three appearance attributes, seven avour or taste attributes
and one attribute, related to viscosity. After development of
the sensory prole, the attributes were evaluated on a 15 cm
non-structured, continuous scale. The left side of the scale (=
0) corresponded to the lowest intensity and the right side of
the scale (= 15) corresponded to the highest intensity of the
attributes. The appearance attribute oily appearance was dened
as the degree of separation into an oily layer as well as the
thickness of this layer. The attribute saturation was dened as
how impenetrable the sample was with respect to light, so that
a very saturated sample was almost impenetrable towards light.
Prior to the sensory evaluation the sensory assessors attended
two training sessions lasting 2 h each for particular training on
milk samples. The assessors were introduced to ve milk samples
which differed signicantly in several of the selected attributes.
The samples consisted of one sample with cardboard avour
and one with stale avour as well as three normal samples
from the feeding experiment. As inspiration the assessors were
introduced to reference samples representing different types
of aroma, avour and taste. The reference samples included
cream as well as milk samples with cardboard avour, stale
avour and metallic avour. These samples were prepared as
described by Hedegaard et al.
5
After each training session the
assessors received feedback on their performance with the aim
of improving and standardising the panels discrimination power.
Due tothe number of samples andthe designof the study, sensory
analyses were carried out on two consecutive days, one type of
concentrate (barley or beet pulp) each day. The effects of type
of concentrate could not be separated from the effect of day of
analysis.
Calculations
Energy corrected milk yield was calculated based on fat and
protein contents.
23
Milk contents of carotenoids and tocopherols
were calculated as content in the milk fat fraction. Based on
analysis of individual feed items the average daily intake of fatty
acids, tocopherols and carotenoids in the TMR were calculated.
Statistical analysis
The effect of treatments (factorial design) on milk yield and milk
composition was analysed by using the GLM procedure of SAS
version 9.1 for Windows (SAS Institute Inc., Cary, NC, USA). As the
full model showedno effects of type of concentrate on production
or milk composition, and as the effects of type of concentrate
could not be separated from the effects of day of analysis for
sensory properties, a decision was made to use a nal model
including the effects of ratio of forage, proportion of concentrate
and their interaction. Multivariate analysis was carried out using
SimcaP+ version 12.0 (Umetrics AB, Umeaa, Sweden) and partial
least squares projections to latent structures was performed using
feed composition as X variables and milk composition as well as
sensory scores as Y variables to establish relationships between
feed composition, milk composition and sensory characteristics.
RESULTS
Feed intake and milk production
Univariate statistical analysis showedthat the milk productionwas
affected by the ratio between maize and lucerne silage and the
proportion of concentrate, whereas the type of concentrate did
not affect the milk composition. Additionally, no signicant effects
of interactionwere observed(Table 2). There was a non-signicant
difference in intake of TMR due to ratio of silage types as well
as proportion of concentrate. The ad libitum intake of TMR was
higher than planned, resulting in a proportion of concentrate of
190 and 360 g kg
1
of dry matter intake, respectively, which was
marginally lower than planned. Milk yield and milk fat and protein
contents increased with proportion of concentrate.
J Sci Food Agric 2013; 93: 9399 c 2012 Society of Chemical Industry wileyonlinelibrary.com/jsfa
9
6
www.soci.org MK Larsen et al.
Table 3. Effect of ratio between maize and lucerne and share of concentrate on milk composition
Proportion of silage,
maizelucerne
Amount of concentrate in
dry matter intake
Effect of silage Effect of amount
Parameter 5 : 1 2 : 1 proportion 20% 40% of concentrate
Fatty acids (mg g fatty acids
1
)
Sum C6C14 19.9 20.0 NS 19.9 20.0 NS
C16 : 0 27.6 29.2 <0.05 31.5 25.2 <0.001
C16 : 1 1.53 1.53 NS 1.64 1.42 <0.001
C18 : 0 11.5 12.4 <0.05 11.5 12.4 <0.05
C18 : 1 tr11 3.11 2.69 <0.05 2.33 3.47 <0.001
C18 : 1 c9 20.7 20.3 NS 18.9 22.2 <0.001
C18:2 n6 1.67 1.59 NS 1.46 1.80 <0.001
C18 : 3 n3 0.40 0.46 <0.01 0.40 0.47 <0.001
CLA cis9 trans11 0.95 0.78 <0.01 0.66 1.07 <0.001
Colour
L, white 86.1 86.5 NS 86.9 85.7 NS
A, red 4.05 4.09 NS 4.03 4.11 NS
B, yellow 5.9 6.9 <0.05 7.05 5.77 <0.01
Fat soluble antioxidants (g g milk fat
1
)
-Tocopherol 9.8 10.4 NS 9.7 10.5 NS
-Tocopherol 0.37 0.32 NS 0.27 0.41 <0.001
-Carotene 1.76 2.16 <0.001 2.05 1.87 NS
Lutein 0.11 0.14 <0.001 0.12 0.12 NS
Urate (mg L
1
) 9.7 9.8 NS 9.9 9.6 NS
NS, not signicant (P > 0.05).
Milk composition
Table 3 depicts the effects of the ratio between maize and
lucerne silage and the proportion of concentrate on milk fatty
acid composition, milk colour and milk content of tocopherols,
carotenoids and urate.
Short-chain fatty acids (C4C14) andmilk -tocopherol content
were not affected by feed composition. C16 and C16 : 1 decreased
along with increased proportion of concentrate, all C18 fatty
acids increased when increasing the proportion of concentrate,
while C18 : 0 and C18 : 3 n3 increased at increased forage
shares of lucerne silage and C18 : 1 trans11 and CLA cis9
trans11 increased when increasing forage shares of maize
silage. Milk yellow colour (b values) increased with increased
forage shares of lucerne and decreased at increased amounts
of concentrate. Milk carotenoids increased at increased forage
share of lucerne. Milk -tocopherol content increased along
with increased proportion of concentrate and milk -tocopherol
was not affected signicantly by the treatments. Milk urate
content was not affected by feed composition and levels
were similar to those previously reported.
11
Apparently, further
comprehensive changes infeedcompositionare requiredtoaffect
milk urate content, and possibly, differences are not observed
unless feed differences are so drastic that milk yield is strongly
affected.
13
Sensory properties
Most sensory attributes were not affected signicantly by
feeding, which shows that the differences due to different
feedings were minor (Table 4). The only attributes that differed
in relation to feeding were the appearance attributes, i.e. oily
appearance, yellowness and saturation, and additionally creamy
avour and stale aroma. Increased proportion of concentrate
in feed resulted in milk with a less oily, less yellow and less
saturated appearance, and increasing the shares of lucerne in
forage resulted in milk with less creamy avour and less stale
aroma.
Multivariate analysis
Multivariate statistical analysis was carried out to establish
relationships between feed components, sensory attributes and
milk composition. The rst model included consumed amounts
of individual feed items (except urea and minerals), and feed
content of fatty acids, tocopherols and carotenoids as X variables,
and milk content of fat, protein, fatty acids, tocopherols and
carotenoids as well as milk colour values and sensory scores as
Y variables. The rst two principal components (PC1 and PC2)
explained 0.85 of the X variance (R
2
X), 0.59 of the Y variance
R
2
Y) and 0.24 of the Y variation could be predicted by the
model according to cross validation (Q
2
), which indicated that
too much noise was modelled. The second model was similar
to the rst one except that the variables, red colour (a value),
metal aroma, sweetness, sourness, bitterness, cardboard avour,
metal avour and creaminess, were omitted. These variables
were omitted because less than 0.10 of the variation could be
predicted by cross validation. Furthermore, univariate analysis
showed small, non-signicant differences (Table 3 and Table 4).
The rst two principal components (PC1 and PC2) of the second
model explained 0.85 of the X variance (R
2
X), 0.71 of the Y
variance R
2
Y) and 0.47 of the Y variation could be predicted by
the model according to cross validation (Q
2
). Plots of scores
and loadings of the second model are shown in Fig. 1 and
Fig. 2. Figure 1 shows how samples were separated by amount
of concentrate by the rst PC and by forage composition by the
second PC.
wileyonlinelibrary.com/jsfa c 2012 Society of Chemical Industry J Sci Food Agric 2013; 93: 9399
9
7
Milk quality affected by forage and concentrate www.soci.org
Table 4. Effect of ratio between maize and lucerne and share of concentrate on sensory attributes
Proportion of silage,
maize : lucerne
Amount of concentrate
in dry matter intake
Effect of Effect of amount
Attibutes 5 : 1 2 : 1 silage proportion 20% 40% of concentrate
Aroma attributes
Creamy aroma 4.2 4.6 NS 4.6 4.2 NS
Metallic aroma 2.7 2.3 NS 2.6 2.4 NS
Cowstable aroma 2.3 2.3 NS 2.1 2.5 NS
Stale aroma 4.0 3.0 <0.05 3.0 3.9 <0.05
Appearance attributes
Oily appearance 5.6 6.0 NS 6.6 5.0 <0.001
Yellowness 6.5 7.0 NS 7.5 6.0 <0.001
Saturation 6.8 7.2 NS 7.4 6.6 <0.10
Flavour and taste attributes
Sweetness 4.7 4.4 NS 4.3 4.8 NS
Sourness 3.6 3.2 NS 3.3 3.5 NS
Bitterness 2.1 2.0 NS 2.2 1.9 NS
Creamy avour 6.8 5.7 <0.05 6.0 6.5 NS
Cardboard avour 2.0 1.9 NS 1.9 1.9 NS
Stale avour 3.2 2.5 <0.10 2.9 2.8 NS
Metallic avour 3.1 3.3 NS 3.4 3.0 NS
Viscosity attribute
Creaminess 6.2 6.0 NS 6.1 6.1 NS
NS, not signicant (P > 0.10).
-4.0
-3.0
-2.0
-1.0
0.0
1.0
2.0
3.0
4.0
-4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0
P
C
2
:

(
X
:

0
.
3
4
,

Y
:

0
.
1
8
)
PC1 (X: 0.52, Y: 0.53)
Barley
Barley
Beet pulp
Beet pulp
Barley
Barley
Beet pulp
Beet pulp
Figure 1. Score plot of partial least squares projections to latent structures
(PLS) of feed components as X variables and milk composition as well
as sensory attributes as Y variables. Triangles: 5 : 1 proportion of maize
and lucerne silage. Squares: 2 : 1 proportion of maize and lucerne silage.
Open symbols: 200 g kg
1
concentrate of dry matter intake. Solidsymbols:
400 g kg
1
concentrate of dry matter intake.
DISCUSSION
The partial least squares projections to latent structures loading
plot (Fig. 2) showed high loadings of the maize silage, lucerne
silage and rapeseed cake, whereas the loadings of barley and beet
pulp were low. This is in agreement with the univariate analysis
where no signicant differences between feeding the two types
of concentrate were observed on the milk composition. Loadings
-0.4
-0.3
-0.2
-0.1
-0.0
0.1
0.2
0.3
0.4
-0.4 -0.3 -0.2 -0.1 -0.0 0.1 0.2 0.3 0.4
P
C
2
:

(
X
:

0
.
3
4
,

Y
:

0
.
1
8
)
PC1: (X: 0.52, Y: 0.53)
F Lutein+b-carotene
F a-tocopherol
F g tocopherol
F C16
F C18:2
F C18:3 n3
BARLEY
BEET PULP
RAPE SEED CAKE
MAIZE SILAGE
LUCERNE SILAGE
L colour
b-colour
a-tocopherol
g-tocopherol
b caroten Lutein
fat
C16
C16:1
C18:0
CLA +C18:1 tr11
C18:1c9
C18:2N6Cis
C18:3n3
C6 to C14
creamy aroma
cow stable aroma
stale aroma
oily appearance
yellowness
saturation
creamy flavour
stale flavour
F C18
F C18:1
Figure 2. Loadings plot of partial least squares projections to latent
structures (PLS) of feed components as X variables (open triangles) and
milk compositionas well as sensory attributes as Y variables (solidsquares).
of chemical composition of feed showed that rapeseed cake was
the main source of -tocopherol, C18 : 0, C18 : 1 and C18 : 2. C16
and -tocopherol in feed derived from rapeseed cake as well as
lucerne silage. C18 : 3 derived from lucerne silage and to a lesser
extent rapeseed cake, and the main source of feed carotenoids
was lucerne silage.
Loadings of individual Y variables were in accordance with
the signicant effects reported in Tables 2 and 3. However, high
loadings of milk white colour (L value), -tocopherol and C6C14
were observed, although no signicant differences between
J Sci Food Agric 2013; 93: 9399 c 2012 Society of Chemical Industry wileyonlinelibrary.com/jsfa
9
8
www.soci.org MK Larsen et al.
treatments were observed for these compounds. Furthermore,
the loading vector of creamy avour had high scores of the rst
as well as the second PC although this attribute was not affected
signicantly by the amount of concentrate which was the main
X variable of the rst PC. Thus, several properties which did
not show signicant differences in the univariate analysis were
affected by treatments according to the multivariate analysis and
such different results should be explained by the different natures
of univariate and multivariate analysis.
Effect of feeding on fatty acid composition
The loadings plot (Fig. 2) shows that higher contents of C18 : 3
n3, C18 : 2 n6 and C18 : 1 in milk fat were directly related to feed
content of these components. The content of C18 : 0 in milk fat
was related to the content of all C18 fatty acids in feed instead
of the content of C18 : 0 in feed, due to the fact that a major part
of the unsaturated fatty acids from feed are hydrogenated in the
ruminal metabolism.
10
These ndings conrm that the content of unsaturated C18
fatty acids in milk fat can be increased by increasing the content
of these compounds in feed. Apart from supplying the fatty acids,
a higher fat content in feed also reduces rumen hydrogenation
which gives a higher recovery of C18 : 3 n3 and C18 : 2 n6.
7,24
The
main feed source of C18 : 3 n3 was lucerne silage, and the content
of this fatty acid in milk fat was higher when the proportion of
lucerne in forage was higher. However, an increased proportion of
concentrate also increased the C18 : 3 n3 content in milk fat, which
was due to a combination of C18 : 3 n3 supply fromrapeseed cake
and the fat content of rapeseed cake, giving a protection against
hydrogenation.
The content of C18 : 1 trans11 in milk fat was also closely
related to the rapeseed cake because this fatty acid is a product
of partial hydrogenation in the rumen of polyunsaturated C18
fatty acids.
10
Higher fat content in rapeseed cake may reduce the
nal hydrogenation step and in this way also increase the C18 : 1
trans11 content in milk fat. CLA cis9 trans11 is mainly formed by
udder desaturation of C18 : 1 trans11 and these two fatty acids
were closely related as commonly found.
2
Other research has
reported a higher content of C16 in milk fat from cows fed maize
silage and higher contents of C18 : 0, C18 : 1 and C18 : 3 n3 in
milk fat are observed when cows are fed lucerne silage due to a
higher fat content in the lucerne based feed.
25
Lucerne has also
been reported to create more stable rumen environments where
hydrogenationis more complete.
26
This results ina higher content
of C18 : 0 in milk fat; however, the content of C18 : 3 n3 in milk fat
is still elevated due to the higher feed content of this fatty acid.
26
As shown in Fig. 2, the milk fat content was negatively related
to the feed content of rapeseed cake. This might be due to a slight,
non-signicant milk fat depression caused by the higher content
of unsaturated fat in the rapeseed cake.
2,9
The content of C16 in milk fat was not related to the feed
content of C16 because this fatty acid is also de novo synthesised.
It is evident that a lower supply of fat from feed leads to a higher
de novo synthesis.
9,27
Effect of feeding on antioxidants
Tocopherols andcarotenoids inthe feedwere closely correlatedto
identical compounds in milk (Fig. 2). These ndings conrm that
the milk contents of tocopherols andcarotenoids canbe increased
by increasing the feed content of these compounds. Figure 2 also
reveals that the main feed source of carotenoids was lucerne
silage, which, as well as rapeseed cake, was the major source
of -tocopherol, while rapeseed cake was the most important
source of -tocopherol. The levels of tocopherol and carotenoids
in milk were similar to values in milk fromcows fed maize silage.
12
When feed concentrations of carotenoids and tocopherols are
enhanced, the effects on milk content decreases as only a limited
amount of carotenoids and tocopherols can be secreted in milk on
a daily basis.
15,28
The present study indicated that concentrations
of antioxidants were relatively low, and this explains the positive
relationship between the content of antioxidants in feed and milk.
Effect of feeding on sensory characteristics
Milk yellow colour (b values) were positively affected by higher
lucerne amounts as well as higher forage intake, and to some
extent related to carotenoids; however, more specically related
to fat. Similarly, yellow colour (b values) are reported to be related
to carotenoid content of milk and feed content of grass products,
however, other components such as riboavin also affect the
colour, and therefore, yellow colour (b values) cannot directly
predict -carotene content.
15,17,29,30
The visual sensory attributes (yellowness, saturation and oily
appearance) were closely connected and were related to creamy
aroma, fat content, yellow and white colour (b and L values)
(Fig. 2). These attributes were associated with high forage levels.
In contrast, high concentrate levels were associated with cow
stable aroma. This was not in accordance with previous results
in which milk from grazing cows has a more barny, salty and
grassy avour than milk fromcows fed a total mixed ration, which
has a sweeter and more malty avour.
17,18
Feeding with maize
has led to a higher intensity of cream avour, maize odour and
sweet corn avour, whereas grass silage has led to an increased
intensity in boiled milk avour, sour avour, cardboard avour
and metallic taste.
3
Although grass silage and the attribute maize
avour were not included in the study, the higher intensity of
creamy avour with increased maize silage was in accordance
with previous results.
3
The results show that milk composition including sensory
properties were affected by composition of forage as well as
proportion of forage to concentrate, whereas the different types
of concentrate did not affect milk quality. This knowledge can be
used to develop guidelines for feeding of dairy cows targeted at
production of milk with distinct properties. If such type of milk
is used for production of dairy products, product composition as
well as sensory properties includingappearance couldbe affected.
ACKNOWLEDGEMENT
This project was fundedbytheDanishMinistryof Food, Agriculture
and Fisheries and by the Danish Cattle Federation.
REFERENCES
1 Dewhurst RJ, Shingeld KJ, Lee MRF and Scollan ND, Increasing the
concentrations of benecial polyunsaturated fatty acids in milk
produced by dairy cows in high-forage systems. Anim Feed Sci
Technol 131:168206 (2006).
2 Grummer RR, Effect of feed on the composition of milk-fat. J Dairy Sci
74:32443257 (1991).
3 Mogensen L, Vestergaard JS, Frett e X, Lund P, Weisbjerg MR and
Kristensen T, Effect of toasting eld beans and of grass-clover:
Maizesilageratioonmilk production, milk compositionandsensory
quality of milk. Livest Sci 128:123132 (2010).
wileyonlinelibrary.com/jsfa c 2012 Society of Chemical Industry J Sci Food Agric 2013; 93: 9399
9
9
Milk quality affected by forage and concentrate www.soci.org
4 Moorby JM, Lee MRF, Davies DR, Kim EJ, Nute GR, Ellis NM, et al,
Assessment of dietary ratios of red clover and grass silages on milk
production and milk quality in dairy cows. J Dairy Sci 92:11481160
(2009).
5 Hedegaard RV, Kristensen D, Nielsen JH, Frost MB, Ostdal H,
Hermansen JE, et al, Comparison of descriptive sensory analysis
and chemical analysis for oxidative changes in milk. J Dairy Sci
89:495504 (2006).
6 Hu FB, Stampfer MJ, Manson JE, Ascherio A, Colditz GA, Speizer FE,
et al, Dietary saturated fats and their food sources in relation
to the risk of coronary heart disease in women. Am J Clin Nutr
70:10011008 (1999).
7 Dewhurst RJ, Fisher WJ, Tweed JKS and Wilkins RJ, Comparison of
grass and legume silages for milk production. 1. Production
responses with different levels of concentrate. J Dairy Sci
86:25982611 (2003).
8 Dhiman TR, Anand GR, Satter LD and Pariza MW, Conjugated linoleic
acid content of milk from cows fed different diets. J Dairy Sci
82:21462156 (1999).
9 Agenas S, Holtenius K, Griinari M and Burstedt E, Effects of turnout to
pasture and dietary fat supplementation on milk fat composition
and conjugated linoleic acid in dairy cows. Acta Agric Scand A Anim
Sci 52:2533 (2002).
10 Harfoot CG and Hazlewood GP, Lipid metabolism in the rumen, in
The Rumen Microbial Ecosystem, ed. by Hobson PN and Stewart CS.
Chapman & Hall, London, pp. 382426 (1997).
11 Havemose MS, Weisbjerg MR, Bredie WLP, Poulsen HDandNielsen JH,
Oxidative stability of milk inuenced by fatty acids, antioxidants,
and copper derived from feed. J Dairy Sci 89:19701980 (2006).
12 Havemose MS, Weisbjerg MR, Bredie WLP and Nielsen JH, Inuence of
feeding different types of roughage on the oxidative stability of
milk. Int Dairy J 14:563570 (2004).
13 Ostdal BH, Weisbjerg MR, Skibsted LH and Nielsen JH, Protection
against photooxidation of milk by high urate content.
Milchwissenschaft 63:119122 (2008).
14 Smet K, De Block J, De Campeneere S, De Brabander D, Herman L,
Raes K, et al, Oxidative stability of UHT milk as inuenced by fatty
acid composition and packaging. Int Dairy J 19:372379 (2009).
15 Noziere P, Graulet B, Lucas A, Martin B, Grolier P and Doreau M,
Carotenoids for ruminants: From forages to dairy products. Anim
Feed Sci Technol 131:418450 (2006).
16 Carpino S, Mallia S, La Terra S, Melilli C, Licitra G, Acree TE, et al,
Composition and aroma compounds of Ragusano cheese: Native
pasture and total mixed rations. J Dairy Sci 87:816830 (2004).
17 Croissant AE, Washburn SP, Dean LL and Drake MA, Chemical
properties andconsumer perceptionof uidmilkfromconventional
and pasture-based production systems. J Dairy Sci 90:49424953
(2007).
18 Khanal RC, Dhiman TR, Ure AL, Brennand CP, Boman RL and
McMahon DJ, Consumer acceptability of conjugated linoleic acid-
enriched milk and cheddar cheese from cows grazing on pasture.
J Dairy Sci 88:18371847 (2005).
19 Frandsen LW, Dijksterhuis G, Brockhoff P, Nielsen JH and Martens M,
Subtle differences in milk: Comparison of an analytical and an
affective test. Food Qual Pref 14:515526 (2003).
20 Stegeman GA, Baer RJ, Schingoethe DJ and Casper DP, Composition
and avor of milk and butter fromcows fed unsaturated dietary-fat
and receiving bovine somatotropin. J Dairy Sci 75:962970 (1992).
21 Slots T, Butler G, Leifert C, Kristensen T, Skibsted LH and Nielsen JH,
Potentials to differentiate milk composition by different feeding
strategies. J Dairy Sci 92:20572066 (2009).
22 Palmquist DL and Jenkins TC, Challenges with fats and fatty acid
methods. J AnimSci 81:32503254 (2003).
23 Sjaunja LO, Baevre L, Junkkarinen L, Pedersen J and Setala J, A Nordic
Proposal for an Energy Corrected Milk (ECM) Formula, EEAP
publication n

50. European Association for Animal Production,
Rome, pp. 156157 (1991).
24 Jiang J, Bjoerck L, Fonden R and Emanuelson M, Occurrence of
conjugated cis-9,trans-11-octadecadienoic acid in bovine milk:
Effects of feed and dietary regimen. J Dairy Sci 79:438445 (1996).
25 Benchaar C, Petit HV, Berthiaume R, Ouellet DR, Chiquette J and
Chouinard PY, Effects of essential oils on digestion, ruminal
fermentation, rumen microbial populations, milk production, and
milk composition in dairy cows fed alfalfa silage or corn silage.
J Dairy Sci 90:886897 (2007).
26 Onetti SG, Shaver RD, McGuire MA, Palmquist DL and Grummer RR,
Effect of supplemental tallow on performance of dairy cows fed
diets with different corn silage: Alfalfa silage ratios. J Dairy Sci
85:632641 (2002).
27 Steinshamn H and Thuen E, White or red clovergrass silage in
organic dairy milk production: Grassland productivity and milk
production responses with different levels of concentrate. Livest Sci
119:202215 (2008).
28 Al-Mabruk RM, Beck NFG and Dewhurst RJ, Effects of silage species
and supplemental vitamin E on the oxidative stability of milk.
J Dairy Sci 87:406412 (2004).
29 Kneifel W, Ulberth F and Schaffer E, Tristimulus color reectance
measurement of milk and dairy-products. Lait 72:383391 (1992).
30 Agabriel C, Cornu A, Journal C, Sibra C, Grolier P and Martin B, Tanker
milk variability according to farm feeding practices: Vitamins A
and E, carotenoids, color, and terpenoids. J Dairy Sci 90:48844896
(2007).
J Sci Food Agric 2013; 93: 9399 c 2012 Society of Chemical Industry wileyonlinelibrary.com/jsfa