COMPONENT TESTING: FAT, PROTEIN AND SOLIDS-NOT-FAT
DAIRY HERD IMPROVEMENT
1985
CASSELL, B.G.
VOLUME: NCDHIP HANDBOOK
The commercial value of milk historically has been influenced by
milk components. In the early years of production testing, when
milk was marketed in the form of butter, the importance of
butterfat to any dairy producer's income was overwhelming.
Consequently, one of the major advances in cowside recording of
production data was the development of the Babcock test for
determining the fat content of milk. As scientific advancement
produced ways to determine protein content and processing plants
began to observe greater returns to cheese manufacturing from
higher protein content in milk, protein also became a more
important milk component. Finally, with the development of rapid,
accurate, automated methods of determining fat, protein and
solids-not-fat (SNF), ever increasing numbers of dairy producers
have had access to such component information on cows in their
herds.
The use of component information for within-herd management
requires that the dairy producer understand the economic importance
of those components to the profitability of the dairy farm.
Breeders of purebred cattle see immediate economic return from the
sale of cattle with pedigree promise for or actual production of
higher milk components. From this view, improved fat, protein or
SNF percentage is economically beneficial. For both purebred and
commercial dairy producers, however, most farm income depends on
the sale of milk. Such returns are based on volume under present
pricing schemes. Breeding programs and within-herd management
practices must recognize that the volume of components produced
rather than the percentage determine income. This publication
explains sources of variation in component percentages and genetic
relationships between components, and mentions sire evaluations for
components. However, the emphasis on component percentages must not
be interpreted as indicating that milk component percentages are
more important than component volume in determining dairy farm
income.
Milk Component Testing in the United States
As of January 1983, approximately 43 percent of all cows in the
United States were on some form of milk recording system (Norman,
Myers and Dickinson, 1983). Some of these plans include component
testing under the supervision of an independent tester. These
plans, along with the 1983 participation in each, are in Table 1.
All cows enrolled in these so-called official plans are tested for
butterfat. The two major divisions of official plans are Dairy Herd
Improvement (DHI) and Dairy Herd Improvement Registry (DHIR).
Within each of these two, dairy producers can participate in AM-PM
testing where weights and samples are obtained from morning or
evening milkings only. To be official, these tests have required
that the milking interval from previous milking be recorded. The
other plan available involves milk weights at both morning and
evening milkings, but component determinations use a milk sample
from only one milking. As Table 1 shows, DHI is the clear leader in
participation in official plans, with DHIR and DHI with AM-PM
component sampling also involving substantial numbers of cows.
Table 2 shows the number of cow years involved in both official and
unofficial testing plans in 1982. While fat testing involves by far
the most cows, testing for protein has grown dramatically in recent
years. In fact, several states initiating protein testing in 1982
or 1983 do not contribute to the nearly 780,000 cow years tested
for protein in 1982. SNF is not a component for which testing is
widespread. Most SNF records originate from the west coast where
SNF plays a role in milk pricing. The difference between cow years
tested for milk and cow years tested for fat under official plans
is a peculiarity of the way these data were collected. All official
plans involve testing for fat percentage.
TABLE 1. Participation in official testing plans as
of January 1, 1983.
Kind of plan National enrollment
Herds Cows
Official DHI 30,829 2,481,408
Official DHI, AM-PM 1,625 137,705
Official DHI, alternate
AM-PM component
sampling 3,189 362,197
Official DHIR 5,248 431,682
Official DHIR, AM-PM 5 682
Official DHIR, alternate
AM-PM component
sampling 133 17,833
Total 41,029 3,431,507
Source: (Norman, Myers and Dickinson, 1983)
TABLE 2. Participation in component sampling
plans for 1982. (a)
Official Nonofficial
cow-years average cow-years average
Milk 3,182,230 15,274 1,215,130 14,138
Fat 3,175,552 3.69 1,143,728 3.76
Protein 779,173 3.27 350,349 3.17
Solids-not-
fat (b) 238,827 8.83
(a) All breeds are combined in averages.
(b) Official and nonofficial plans combined.
Source: (King and Wiggans, 1983)
Table 2 shows that participation in component testing under plans
outside the official designation is widespread. Such numbers
demonstrate the importance of component test results to
participating dairy producers for within-herd management.
Sources of Variation
Component tests vary from cow to cow, from lactation to lactation
with each individual cow and from test day to test day for a given
cow. Such variation is a natural part of the biological process of
making milk components. Age, stage of lactation, month of
freshening, illness, injury, feed, reproductive processes, climate,
milking procedures and equipment, sampling techniques, sample
shipment and lab procedures are possible sources of differences in
component test results. Some of these factors operate in a rather
random fashion, in some cases increasing and in others decreasing
component test results. Some of the factors may be systematic in
nature, introducing bias into milk component test results. Further,
fat tests may respond quite differently from protein under the
influence of some factors.
The time during milking when a sample is drawn is known to have a
dramatic effect on the fat test. Gilmore and Gaunt sampled milk
drawn the first minute, second minute, third minute and after 3
minutes of milking on 10 cows in 4 breeds (Gilmore and Gaunt,
1962). Their results are reported in Table 3. Fat test results
increase dramatically from first minute's milk to stripping. On the
other hand, protein and SNF results are affected little by time of
milking when a sample is drawn. These results show the importance
of a representative sample for butterfat determination. All of the
milk produced by a cow at one milking must be agitated thoroughly
before a sample is drawn. Table 3 also supports a clear argument
for milking cows completely at all times, not just on test day.
Time during milking was one factor in which fat and protein behaved
differently. Major environmental effects, however, seem to affect
fat and protein in similar ways. Table 4 shows the effect of season
of test on fat and protein content from cattle in Quebec
(Ng-Kwai-Hang, Hayes, Moxley and Monardes, 1984). The fat test
peaks in winter just as the protein test does. Lowest protein and
fat tests occur in summer. Between the extremes, both fat and
protein tests show the same trends. Such gross effects as season
(or month) of test can represent a combination of many factors
including temperature and feed sources. In general, component tests
go down as temperature goes up and vice versa. Of course, many
other factors could influence component test results besides
temperature. Nutrition has a profound effect, as many dairy
producers who have fed low fiber rations can attest. Some seasonal
differences in dairy rations could be involved.
TABLE 3. Changes in percentage of milk fat, pro-
tein and solids-not-fat during milking. (a)
Milking period Average
Percentage Percentage Percentage
fat protein SNF
1st minute 2.46 3.49 8.97
2nd minute 3.62 3.46 9.07
3rd minute 4.90 3.40 8.98
Over 3 minutes 6.30 3.34 8.89
(a) Data were from 10 cows each in Ayrshire, Guernsey,
Holstein and Jersey breeds.
Source: (Gilmore and Gaunt, 1962)
TABLE 4. Effect of season of test on fat and protein
percentages.
Season Effect
Fat percentage Protein percentage
Winter highest highest
Spring decreasing decreasing
Summer lowest lowest
Fall increasing increasing
Source: (Ng-Kwai-Hang, Hayes, Moxley and Monardes, 1984)
Stage of lactation influences test. Both fat and protein tests
follow a similar trend through the lactation. Component percentages
decline rapidly during the first month of lactation, changing by .5
percent or a bit more for both fat and protein. Lowest tests occur
during the second and third months of lactation. Beginning at about
the fourth month of lactation, component percentages increase in a
linear fashion, reaching a level very near the tests obtained
immediately after freshening by the 10th or 11th month of
lactation. A cow's age also affects fat and protein percentages.
The fat test declines very slightly as a cow ages. The protein test
actually may show an increase from first to second lactation, but
then decline just as the fat test does as a cow increases in age
(Ng-Kwai-Hang, Hayes, Moxley and Monardes, 1984).
Frequency of testing is a factor in determining component
percentages. For many years, monthly testing has been the standard
against which all testing schemes are evaluated. Is there anything
magical about monthly testing? Actually, monthly testing (for 24
hours of milk) is probably oversampling for some production
measures and undersampling for others. Table 5 shows the results of
a study on the accuracy of AM-PM component sampling. In the study,
each milking's milk was sampled and analyzed separately. A
composite sample also was obtained and tested. Notice that PM milk
tests considerably higher for fat than AM milk (Lee and Wardrop,
1984). Unfortunately, the difference is not sufficiently
predictable to allow adjustments to be made. Predicting daily
(24-hour) results from an AM or PM sample gave answers too high by
.06 percent. For individual cows, the predicted results were in
much greater error. For only about 50 percent of the cows, the
predicted results were within .2 percent of the composite sample.
Almost 20 percent of AM predictions missed the composite by.4
percent or more. Such differences greatly compromise fat test
results for within-herd management decisions. If an extremely
valuable cow was involved with a low test, most dairy producers
would be unwilling to accept the results of such a testing scheme.
Notice, however, that AM-PM component sampling is very successful
for protein. AM and PM results are so close to composite test
results that no adjustments are needed. Milk production also can be
accurately determined from AM-PM testing provided that accurate
milk weights are a priority for the dairy producer involved.
TABLE 5. Averages and standard deviations of morning,
evening and composite milk samples.
Standard
Variable Average deviation
Composite fat percentage 3.83 0.45
AM fat percentage 3.60 0.54
PM fat percentage 4.12 0.58
Daily predicted from
AM sample 3.88 0.58
Daily predicted from
PM sample 3.89 0.55
Composite protein percentage 3.35 0.30
AM protein percentage 3.34 0.33
PM protein percentage 3.35 0.30
Source: (Lee and Wardrop, 1984)
Dairy producers participating in AM-PM component sampling programs
should be aware of the results found by Lee and Wardrop as shown in
Table 5. Monthly fat tests that are used for ration formulation and
perhaps culling are not as accurate as they should be for optimum
within-herd management decisions. On the other hand, fat tests for
the entire lactation under AM-PM component sampling are less
subject to error in a single monthly test and are more accurate and
useful than the monthly results. The decision to convert from DHI
or DHIR to an AM-PM component sampling plan should be considered
carefully. Some loss in accuracy of fat test determination will
occur with AM-PM component sampling. Whether such a loss in
accuracy is worthwhile depends on a dairy producer's use of monthly
fat test results, any possible reduction in testing costs,
interruption of milking schedules with sampling from two milkings,
and other considerations. Above all, a dairy producer should not
convert to AM-PM component sampling because the tester or someone
else said it would make no difference in quality of information.
AM-PM component sampling does make a difference, and the dairy
manager is the one who suffers from loss of information.
Changes in fat or protein test for individual cows from month to
month are expected. This is one reason why many management
decisions are based on total lactation results rather than a single
month's test. Month-to-month variation of a random nature is
largely, but not completely, eliminated by examining milk
components on a lactation basis. Wilcox, et al. published
repeatabilities for milk components from one lactation to the next.
These estimates help determine the need for additional records on
an individual cow. The estimates were.76 for fat percentage,.60 for
protein percentage, and.61 for SNF percentage in Holsteins, with
similar results in the other breeds. Repeatabilities for milk and
fat yield in Holsteins were lower at .53 and .49 percent (Wilcox,
Gaunt and Farthing, 1971). In general, milk component percentages
observed in one lactation are very useful predictors of component
percentages in subsequent lactations and are more useful for this
purpose than is milk yield.
Genetic Control of Milk Components
The previous discussion covered nongenetic reasons for monthly
differences in component tests on an individual cow or from cow to
cow. As Table 6 shows, there is an important genetic component in
variation of milk components. These data are the results of a
cooperative study of many research herds in the United States
(Wilcox, Gaunt and Farthing, 1971). Heritability tells us what
percentage of total variation for a trait is the result of genetic
differences between animals. Heritability also is an important
factor in the rate of genetic change. For milk components, the
opportunity for change is great. Heritability estimates are in the
range of .5 for fat percentage, .4 to .5 for protein and .5 or
higher for SNF. For comparison, the heritability of milk or
component yield is only 0.2 to 0.25 percent. If a dairy producer
was to try to change milk composition by selection, chances for
success would be very good. However, as pointed out in the
introduction, product value is determined by volume rather than
percentage. Selection on component percentages generally is not
recommended.
Most dairy producers are well aware that increases in milk
component percentages tend to be associated with reduced milk
volume. Genetic correlations between milk component percentages and
milk yield substantiate this general rule. The genetic correlation
measures the extent to which two traits are controlled by the same
genes. The range of the genetic correlation is -1 to +1. Values
from .4 to 1.0 indicate that two traits will progress strongly in
the same direction if selection is for one of the traits alone.
Genetic correlations around zero indicate that two traits operate
in a fairly independent fashion. Genetic correlations from -.4 to
-1.0 indicate that two traits will diverge rather strongly if
selection is for one trait alone. In Table 7, all genetic
correlations between milk components and milk yield are negative
(Wilcox, Gaunt and Farthing, 1971). Some of the negative
correlations are large enough to indicate rather strong divergence
between test and yield if selection is for one or the other trait
alone. Recall that economic considerations recommend maximizing
volume of component rather than percentage content. Since milk
volume and component percentage are antagonistic, selection could
be based on the economic value of the total product produced by the
cow. This is exactly what Predicted Difference Dollars (PD$),
published by USDA in biannual Sire Summaries, does. Bulls siring
desirable volumes of components appear at the top of the list in
order by PD$.
Readers should be aware that the estimates of heritability and
genetic correlations in Tables 6 and 7 are exactly that, estimates.
The heritability of SNF percentage for Ayrshires of 1.01 in Table
6 is above the theoretical maximum of 1.00. Small data samples can
cause such results. Also, readers should not overinterpret breed
differences in genetic parameters. It is impossible to segregate
true breed differences from sampling error in Tables 6 and 7.
Sire Evaluations for Protein
USDA has published sire evaluations for milk, fat and fat
percentage for many years. Dairy producers rely on USDA sire
evaluations and are familiar with good and bad evaluations for fat
percentage. Protein evaluations are newer and less familiar to
dairy producers. Table 8 contains a summary of evaluations for
protein and fat percentage from the Summer 1984 Modified
Contemporary Comparison. Not all bulls with proofs for milk and fat
have proofs for protein. This is unfortunate for dairy producers
interested in placing some selection emphasis on protein
percentage. It means that some of the top bulls for PD$ or PD milk
may not have protein proofs and dairy producers may pay less
attention to such bulls than their proofs for milk and fat deserve.
Fortunately, the percentage of active AI bulls with protein proofs
has increased dramatically over the past year. At least two out of
three active AI bulls in each breed now have protein proofs and the
percentage continues to increase.
TABLE 6. Heritability estimates for milk compo-
nents by breed.
Breed Component (percentage)
Fat Protein SNF
Ayrshire 0.54 0.35 1.01
Guernsey 0.56 0.49 0.45
Holstein 0.57 0.37 0.54
Jersey 0.71 0.56 0.63
Brown Swiss 0.51 0.69 0.32
Source: (Wilcox, Gaunt and Farthing, 1971)
TABLE 7. Genetic correlations between milk com-
ponents and milk yield by breed.
Breed Genetic Correlation
with milk yield
Percentage Percentage Percentage
fat protein SNF
Ayrshire -.17 -.21 -.09
Guernsey -.35 -.20 -.24
Holstein -.30 -.30 -.22
Jersey -.56 -.55 -.21
Brown Swiss -.39 -.11 -.17
Source: (Wilcox, Gaunt and Farthing, 1971)
Good proofs for protein and fat percentages are not necessarily the
same; protein is less variable than fat. Thus, dairy producers will
not see as much difference between the best and worst bulls for
protein as they will for fat. Table 8 contains some useful numbers
to keep in mind when evaluating a bull's protein or fat proof. For
instance, the top 25 percent of active AI Holstein bulls currently
have proofs of 0.00 or better. The top 10 percent of Holstein bulls
exceed 0.02 for protein percentage. As an example of the
differences in distribution of protein and fat proofs, the top 25
percent of active AI Holsteins exceed.05 for fat percentage. The
top 10 percent for fat percentage exceed .12.
TABLE 8. Sire evaluations for protein in Summer
1984.
Percentage Lower limit
Breed No. active of all Average Top Top
sires with active proof 25% 10%
proof sires (percentage)
protein
Ayrshire 14 87.5 -0.03 -0.01 0.05
Guernsey 24 75.0 -0.05 -0.03 0.00
Holstein 329 67.3 -0.03 0.00 0.02
Jersey 48 85.7 -0.06 -0.02 0.03
Brown Swiss 26 72.2 -0.02 0.02 0.04
fat
Ayrshire 16 100 -0.03 0.03 0.12
Guernsey 32 100 -0.06 0.02 0.12
Holstein 489 100 -0.01 0.06 0.12
Jersey 56 100 -0.03 0.08 0.14
Brown Swiss 36 100 0.01 0.09 0.16
Source: Author's investigation
Summary
Milk components have played an important role in production testing
programs since the advent of such programs shortly after the turn
of the century. The usefulness of milk component data for
within-herd management decisions will increase as more and more
dairy managers incorporate such data into ration calculations,
culling decisions and even into monitoring herd health. Continued
interest in component pricing schemes seems likely as consumer
preferences continue to shift from animal fats toward higher
protein products. Variations in milk pricing systems from one area
of the country to another will complicate sire selection, making
indexes using regional component prices more appealing.
The interest in milk component determinations, currently prevalent
in the dairy industry, is likely to continue in the near future.
Dairy herd managers and those responsible for production testing
programs in the states need to be aware of sources of variation in
component test results. Such knowledge will improve the
interpretation and help make the use of component data more
beneficial to the individual herd and the dairy industry as a
whole.
References
1. Gilmore, H.C. and S.N. Gaunt. 1962. Changes in percentage of
protein, milk fat, and solids-not-fat during the milking process.
J. Dairy Sci. 45:1574.
2. King, G.J. and G.R. Wiggans. 1983. USDA summary of 1982 U.S. cow
herd averages. Dairy Herd Improvement Letter, Vol 59, No. 5. pages
1-24.
3. Lee, A.J. and J. Wardrop. 1984. Predicting daily milk yield, fat
percent, and protein percent from morning or afternoon tests. J.
Dairy Sci. 67:351.
4. Ng-Kwai-Hang, K.F., J.F. Hayes, J.E. Moxley and H.G. Monardes.
1984. Variability of testday milk production and composition and
relation of somatic cell counts with yield and compositional
changes of bovine milk. J. Dairy Sci. 67:361.
5. Norman, H.D., E.F. Myers and F.N. Dickinson. 1983. National
Cooperative Dairy Herd Improvement Program participation
report--state activities as of January 1, 1983. Dairy Herd
Improvement Letter, Vol 59, No. 3.
6. Wilcox, C.J., S.N. Gaunt and B.R. Farthing. 1971. Genetic
interrelationships of milk composition and yield. Southern
Cooperative Series, Bulletin No. 155.
a
תתתתתתתתתתתתתתתתתתתתThe National Dairy Database (1992)תתתתתתתתתתתתתתתתתתתת
תתתתתתתתתתתתתתתתתתתתתתת\NDB\DAIRY\TEXT\DA101200.TXTתתתתתתתתתתתתתתתתתתתתתתת
%f TITLE;COMPONENT TESTING: FAT, PROTEIN AND SOLIDS-NOT-FAT
%f COLLECTION;DAIRY HERD IMPROVEMENT
%f ORIGIN;Virginia
%f DATE_INCLUDED;June 1992