FOOD COMPOSITION AND ADDITIVES

Comparison of Babcock and Ether Extraction Methods for

Determination of Fat Content of Cream: Collaborative Study
LYNCH ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 79, NO. 4, 1996
JOANNA M. LYNCH and DAVID M. BARBANO
Cornell University, Northeast Dairy Foods Research Center, Ithaca, NY 14853
J. RICHARD FLEMING
U.S. Department of Agriculture, Texas Milk Marketing Service, Carrollton, TX 75006

A modified Mojonnier ether extraction method for

determination of the fat content of cream was developed based on the method for milk (AOAC Official Method 989.05). The cream Babcock method
(AOAC Official Method 920.111BC) was modified
to harmonize with the milk Babcock method
(AOAC Official Method 989.04) and to clarify procedural details. Using the AOAC collaborative study
format, 10 laboratories tested 9 pairs of blind duplicate heat-treated cream samples with a fat range of
3045% using both methods. The statistical performance (invalid and outlier data removed) was as
follows: mean % fat = 37.932, sr = 0.125, sR = 0.151,
RSDr = 0.330, RSDR = 0.398, r = 0.354, and R =
0.427 for the ether extraction method. For the Babcock method, mean % fat = 38.209, sr = 0.209, sR =
0.272, RSDr = 0.548, RSDR = 0.712, r = 0.592, and R
= 0.769. Average test results for fat from the Babcock method were 0.277% (absolute fat) greater
than for the Mojonnier ether extraction method.
The difference between methods, as a percentage
of the average fat content of the samples, was
0.73%. This agrees with differences observed between the 2 methods for milk when 10 to 17 laboratories tested 7 milk samples in blind duplicate at bimonthly intervals over a 4-year period (average difference 0.029% fat, 0.78% as a percentage of
average fat content). The Mojonnier ether extraction and Babcock methods for fat in cream have
been adopted by AOAC INTERNATIONAL. The new
Babcock method replaced the AOAC Official
Method 920.111BC.

Submitted for publication May 25, 1995.

The recommendation was approved by the Methods Committee on
Commodity Foods and Commodity Products and was adopted by the
Official Methods Board of the Association. See Official Methods Board
Actions (1995) J. AOAC Int. 78, 179A, and Official Methods Board
Actions (1995) The Referee, October issue.

n the cream trade, the fat test is used as a basis for payment.
The financial implications of fat testing to the buyers and
sellers of cream underscores the importance of having accurate and precise testing methods. The Babcock test is the
most common method for determination of fat content in
cream, although the performance characteristics of the Babcock method for cream (AOAC Official Method 920.111BC;
1) are currently not defined. The modified Mojonnier ether extraction test for milk (AOAC Official Method 989.05; 2) is also
used to test cream, although technically the AOAC method
does not currently apply to cream. Cream fat testing by the
Rose-Gottlieb method, which utilizes the basic principle of
ether extraction, is a standard International Dairy Federation
(IDF) method (3).
For raw milk, however, the performance characteristics of
the Babcock (AOAC Official Method 989.04; 4) and ether extraction (AOAC Official Method 989.05; 2) methods were described in 1988 (5), and the methods were adopted by AOAC
in 1989. For the milk Babcock test, the method collaboratively
studied in 1988 incorporated improvements and clarifications
of the original 1920 method (5). Many of the improvements
and clarifications could apply to the cream Babcock method
(1), which does not currently incorporate these changes. The
ether extraction method for milk (2) has the potential to be used
for cream, with modifications to describe differences in sample
preparation between cream and milk.
In the 1988 milk collaborative studies (5), it was observed
that the Babcock method gave higher test results than the ether
extraction method (average 0.021% fat). Over a 4-year period,
10 to 17 laboratories tested 7 milk samples in blind duplicate at
bimonthly intervals; the difference between methods was
0.029% fat (6). This latter estimate is probably more representative of the actual difference between methods, as method
performance during the 4-year period was superior and was
derived from more laboratories than was the original
1988 study. The difference in test results between the Babcock
and ether extraction methods for cream has not been defined,
although it would be expected to be similar relative to fat content.
The objectives of the present study were to improve and
clearly define the procedural details of both the Babcock and
Mojonnier ether extraction methods, define method perform-

ance, and quantitate the difference in the estimation of cream

fat content between the 2 methods.
Preliminary Work: Babcock Method

Methods 1 and 2
The cream Babcock method (1) currently describes
2 method options (methods 1 and 2), 3 types of bottles (50%
9 g short neck, 50% 9 g long neck, and 50% 18 g long neck); it
does not address a fat range within which the method is applicable. Method 1 can be used for all 3 types of bottles whereas
method 2 can be used with 50% 9 g bottles only. Ideally, a single method applicable to all types of bottles is most desirable.
The differences between methods 1 and 2 involve fairly minor procedural details. In method 1 (for 50% 9 g bottles), about
9 mL sulfuric acid is added to 9 g cream and a reaction temperature between 93 and 103C is achieved. This is followed
by the addition of 5 to 10 mL water. The final column appearance is golden amber. In method 2, 9 mL water is added to 9 g
cream followed by the addition of about 17.5 mL sulfuric acid.
A reaction temperature of about 108C is achieved and the final
column appearance is golden. Undigested material may be present in the fat column unless the samples are shaken between
the first and second centrifugation. Extensive in-house testing
and sample exchanges between 2 different laboratories indicated no difference in test results (absolute test result or repeatability) between methods 1 and 2.

within a relatively narrow range. With cream, however, water

content varies widely.
Reaction temperatures were determined for samples with
3045% fat, and ranged from 93 to 103C depending on the
fat content. Lower fat samples had lighter column colors and
higher reaction temperatures than did higher fat samples. High
fat cream samples could get quite dark (almost brown) with no
char, whereas lower fat samples show char at lighter column
colors. The amount of sulfuric acid added (7 to 10 mL) affected
column appearance but not reaction temperature. This is an important point, as the amount of sulfuric acid is used to manipulate reaction temperature in the milk Babcock method. In the
cream method, however, the amount of sulfuric acid to be
added was also determined by final column appearance.
The amount of water to be added prior to centrifugation was
changed from the current AOAC specifications of 510 mL
soft H2O at 60C or above to 6 mL distilled H2O at 56.5
61C. The volume was specified as a single addition (6 mL)
to simplify the procedure, as many laboratories try to manipulate the water addition to achieve good columns when a range
is given. The 6 mL addition worked well for both 9 g and 18 g
bottles. A liberal temperature range of 56.661C was selected
in response to comments from the participating laboratories. At
this temperature range, water can be tempered either in a selfcontained water reservoir set at 5961C or in the Babcock
tempering bath set at 56.558.5C without affecting test results or column appearance.

Selection of Method
Bottle Specification
Although methods 1 and 2 are similar, it is unnecessarily
confusing to have 2 methods for testing cream that differ in
relatively minor procedural details. Between-laboratory reproducibility is best when a single standardized method is used, as
it minimizes procedural differences among laboratories.
Method 1, which has some advantages over method 2, was selected for the collaborative study. It can be used with all types
of 50% bottles, whereas method 2 can be used only with 50%
9 g bottles, because the bulb of the 18 g bottles (about 50 mL
volume) cannot accommodate the volumes necessary to conduct method 2 with 18 g cream (18 g cream, 18 mL water,
35 mL sulfuric acid). Method 1 also uses less acid than
method 2, and thus has a cost advantage. Finally, unlike
method 2, method 1 can achieve good columns (no undigested
material) without shaking between the first and second centrifugations.

Method Development
Method 1 (1) was modified to harmonize with the current
milk Babcock method (4) and to clarify the procedural details
specific to testing cream. Most important was specification of
reaction temperature and column appearance. Reaction temperature and column appearance are dependent on the water
content of the samples: the greater the water content, the higher
the reaction temperature and the lighter the columns. In the
milk Babcock method, a single reaction temperature of 108C
is specified because the water content of milk samples falls

The proposed specifications of the cream Babcock bottles

were written to harmonize with those for milk Babcock bottles
(4). In addition, we took into account the current AOAC cream
bottle specifications (1), the cream Babcock bottles already in
existence, and the accuracy of the apparatus (Nafis testers and
modified Nafis testers) for testing the bottles (7). We also
sought input from the various Babcock bottle manufacturers.
Three types of Babcock bottles were described: 50% 9 g
short neck, 50% 9 g long neck, and 50% 18 g long neck. Short
neck cream bottles are similar in design to 8% milk bottles; the
graduated portion of the neck is specified as 75 mm (4). Long
neck bottles have a longer scale, 120 mm, and a potential for
greater accuracy and readability (1). However, because many
laboratories can use only short neck bottles, 9 g short neck bottles used in the collaborative study. They are the most common
bottles used, they fit all types of Babcock centrifuges, and their
performance represents minimum achievable repeatability and
reproducibility.
Other Babcock bottles exist for different products such as
20% or 30% bottles for measuring the fat content of light
cream. However, it was not within the scope of this study to
investigate all fat levels and all types of bottles. The 20 and 30%
bottles have the potential for better repeatability and reproducibility for measuring the fat content of light cream, and thus are
commonly used for this purpose.

Applicability
Normal medium to heavy cream ranges in fat content from
30 to 45%. The Babcock test is suitable for use with raw, heattreated, or pasteurized cream. It cannot be used with homogenized products such as homogenized light cream, because the
fat cannot be cleanly extracted.
Preliminary Work: Ether Extraction Method

Method Development
The modified Mojonnier ether extraction method for milk
specifies a sample size of 10 g (2). However, this sample size
cannot be used for cream because of problems with emulsification. Similarly, severe emulsification problems are observed
with the Rose-Gottlieb method (AOAC Official
Method 920.111A) for cream (8), which specifies a sample size
of 5 g.
The IDF Rose-Gottlieb method for cream specifies a sample
size yielding 0.30.6 g extracted fat (3), which is similar to the
amount extracted from 10 g milk (2). This is followed by the
addition of 50C water to a total volume of 1011 mL (3). This
sample size and water addition was adopted for use with the
modified Mojonnier ether extraction method for cream. Because no difference resulted from using 50C or room temperature water, a water temperature of 2022C (room temperature) was specified.

Applicability
The ether extraction test is theoretically applicable to cream
at all fat levels. In the collaborative study, cream with a fat
range of 30 to 45% was tested so that the results could be directly compared with those of the Babcock method. A single
sample size of 1 g was specified for the collaborative study. The
ether extraction test performed equally well with pasteurized,
heat-treated, homogenized, or raw cream.
Collaborative Study
Babcock bottles (50% 9 g short neck) for the collaborative
study were purchased by each laboratory. The volume of the
graduated portion of the neck of the bottles was checked at a
central laboratory using a modified Nafis testing apparatus (7).
Only bottles that met the proposed specifications for 50% 9 g
short neck bottles were used.
Each of 10 laboratories received 2 sets of 18 cream samples
(9 pairs of blind duplicates). One set of samples was used for
the Babcock test and the other was used for the ether extraction
test. Samples were coded with random 3-digit numbers. The
codes were different between and within sets, and analysis order was randomized for all laboratories and methods.
Nine batches of pasteurized cream and the corresponding
separated pasteurized skim milks were obtained from 6 different locations throughout the continental United States and
shipped on ice by overnight mail to a central location. Each of
the 9 creams was collected at a different commercial dairy
plant. To obtain a sample fat range of 3045%, 3 creams (num-

bers 2, 3, and 4) were diluted with their corresponding skim to

achieve the desired fat content.
Creams were collected on Wednesday or Thursday and arrived at the central laboratory on Friday. The cream was coldsplit on Monday. The cream from each dairy plant was mixed,
poured into 1 large plastic container, and agitated continuously
with a motor driven stirrer while the cream (about 100 mL) was
drawn from a spout directly into pre-coded 6 oz Whirl-Pak
bags (Nasco, Fort Atkinson, WI). Samples were refrigerated
immediately after splitting. Splitting uniformity was verified
by checking the fat content of the first, middle, and last sample
bag using a Dairy Laboratory 2 mid-infrared milk analyzer
(Multispec, Ltd, York, UK).
The sample bags, wrapped in plastic bags and sealed with
tape to protect them from water contamination, were packed on
wet ice and shipped to the individual laboratories by overnight
air delivery. Samples arrived at the laboratories on Tuesday.
Arrival temperature of the creams was verified as 6C by each
laboratory.
Forms were provided for recording raw data, test results,
and comments about individual samples. A questionnaire was
provided for each test method. Information such as sample arrival time and actual testing conditions during analysis, was
requested to ensure that analysts followed all details of the procedures. All chemical tests were conducted within 3 days of
receipt and results were sent back to the Associate Referee.
Data were analyzed using the AOAC guidelines for calculating
statistical parameters and identifying outliers (9).
995.18 Fat in Cream, Babcock Method
Final Action
Revised First Action 1995
(Applicable to raw and heat-treated cream containing 30
45% fat.)
(Caution: See Appendix B, safety notes on centrifuges, sulfuric acid, and mercury. Dispose of waste solvents according
to applicable environmental rules and regulations.)
Method Performance:
sr = 0.209; sR = 0.272; RSDr = 0.548%; RSDR = 0.712%; r
= 0.592; R = 0.769.

A. Principle
Known weight of cream is placed into Babcock cream bottle
and H2SO4 is added. Generated heat releases fat, which is then
isolated by centrifuging and addition of H2O. Fat is quantitated
in graduated portion of Babcock bottle. Result is expressed as
% fat by weight.

B. Reagents
(a) Sulfuric acid.Specific gravity 1.821.83 at 20C.
(b) Red mineral oil.White mineral oil containing red
oil-sol artificial color; specific gravity <0.85 at 20C.
(c) Mercury.Purity 99.99%. Used for testing accuracy
of Babcock bottles.

C. Apparatus
(a) Test bottles.Standard Babcock cream-test bottles,
50%, 9 g, either short- or long-neck, or 50%, 18 g, long-neck;
resistant to stress created during centrifugation. Bottles must
have mark on top of neck above graduations, indicating weight
of sample to be used (i.e., 9 or 18 g). Standard Babcock creamtest bottle specifications are as follows:
(1) 50%, 9 g, short-neck, standard cream-test bottle.Total height 155170 mm (6.16.7 in.). Bottom of bottle must be
flat, and axis of neck vertical when bottle stands on level surface. Quantity of cream for bottle is 9 g.
Bulb.Capacity of bulb to junction with neck is 45 mL.
Bulb may be either cylindrical or conical. If cylindrical, od
must be 3436 mm; if conical, od of base must be 3133 mm,
and maximum diameter, 3537 mm.
Neck.Cylindrical and of uniform diameter from 5 mm
below lowest graduation mark to 5 mm above highest. Top of
neck is flared to diameter of 15 mm. Graduated portion of
neck is 75 mm long and is graduated in 5%, 1%, and 0.5%
from 0.0 to 50%. Graduations may be etched, with black or
dark pigment annealed to graduation, or may be unetched black
or dark lines permanently annealed to glass. Graduation line
widths must be 0.2 mm. The 0.5% graduations must be 4
6 mm long; 1% graduations 6 mm long, and must project
2 mm to left of 0.5% graduations. The 5% graduations must
extend at least half-way around neck to right and project
4 mm to left of 0.5% graduations. Each 5% graduation must
be numbered (thus: 0, 5, 10, ... 45, 50), with numeral placed
either to left of 5% graduations, or 1 mm to left of 1% graduations and 1 mm above the 5% graduations. Numbers should
be placed in such way that they are visible during measurement
of fat column.
Vertical line must be etched and annealed with black or dark
pigment or must be etched into dark stripe permanently annealed to glass. Etched vertical line must be located 5.5
8.5 mm to right of the left end of 0.5% graduation marks and
extend 1 mm above 50% line and 1 mm below 0% line. If
vertical line is etched into dark vertical stripe, then dark vertical
stripe must extend 1 mm above the 50% line and 1 mm below 0% line. Dark vertical stripe should be 2.0 mm wide and
etched line must be 0.2 mm wide.
Zero line (0.2 mm wide) must be etched and annealed
with black or dark pigment, or not darkened but etched through
a dark horizontal stripe (2.0 mm wide) permanently annealed
to glass. Zero and vertical etched lines, with or without annealed pigment, should be of sufficient depth to hold point of
dividers.
Capacity of neck for each whole % on scale is 0.100 mL.
Testing.Accuracy of each bottle shall be determined before use (usually by manufacture or certified laboratory). Bottle
calibration accuracy is determined by placing bottle upside
down on Babcock bottle calibration apparatus (Nafis tester or
modified Nafis tester) that is capable of delivering known volumes of Hg into Babcock bottle neck. Bottle calibration apparatus delivery is calibrated and volume of Hg contained between 50 and 25% (2.500 mL), 25 and 0% (2.500 mL), and 50

and 0% (5.000 mL) marks is determined. Accuracy of bottle

neck calibration must be determined before or after attachment
to bulb. Maximum error of total graduations or any part thereof
must not exceed 0.050 mL.
(2) 50%, 9 g, long-neck, standard cream-test bottle.
Same as in (1), except total height must be 210229 mm (8.25
9.0 in.) and graduated portion of neck must have length
120 mm.
(3) 50%, 18 g, long-neck, standard cream-test bottle.
Same as in (1), except for following: total height is 210
229 mm (8.259.0 in.), quantity of cream for bottle is 18 g,
length of graduated portion of neck is 120 mm, and capacity
of neck for each whole % on scale is 0.200 mL. Perform testing
of bottles as in (1), except that volume of mercury contained
between marks is twice of that of 50%, 9 g, short-neck, standard cream bottles.
(b) Acid measure. Device used to measure H2SO4 should
deliver in range 515 mL, if 9 g bottles are used, and 10
20 mL, if 18 g bottles are used. Use device that can be set to
consistently deliver appropriate amount of acid to obtain desired cream acid reaction temperature in Determination, D.
(c) Centrifuge or tester.Standard centrifuge, however
driven, must be constructed throughout and so mounted as to
be capable, when filled to capacity, of rotating at necessary
speed with minimum vibration and without liability of causing
injury or accident. Centrifuge must be heated, electric or otherwise, to 5560C during centrifuging. It must be provided
with speed indicator, permanently attached, if possible. Proper
rate of rotation may be determined by reference to Table 989.04
(4). Rotation speed with full centrifuge should be checked periodically with tachometer. Diameter of wheel means distance between inside bottoms of opposite cups measured
through center of rotation of centrifuge wheel while cups are
horizontally extended. Dimensions of centrifuge or tester must
be compatible with type of cream-test bottle used (i.e., short- or
long-neck).
(d) Dividers or calipers.For measuring fat column.
(e) Water bath for test bottles.Provided with thermometer and device to maintain temperature of fat column at 57.5
1C. Dimensions of H2O bath must be compatible with the type
of cream-test bottle used (i.e., short- or long-neck).
(f) Water bath for tempering cream samples prior to weighing.Provided with thermometer and device to maintain temperature of cream at 38 1C.
(g) Supply of hot water added after 1st and 2nd centrifugations.Provided with thermometer, device to maintain temperature of distilled H2O at 5961C, and device to deliver
H2O into Babcock bottles.
(h) Bottle shaker.Variable speed and matched to maximum capacity of centrifuge.
(i) Digital thermometer for measurement of creamacid reaction temperature.Capable of reading temperature to the
nearest degree in range of 85105C. Use acid-resistant probe,
with small diameter (0.5 mm) to ensure rapid response time.
Length of temperature probe should be such that its tip is ca
1 cm above bottom of bottle, when fully inserted.

(j) Reading light.As background when measuring fat

columns. Light should be diffused (soft white color) and provide illumination from angles above and below level of fat column. Magnification device must be used to aid reading.
(k) Analytical balance.Capable of weighing to the nearest 0.01 g. Check accuracy periodically and whenever balance
is moved or cleaned. Keep record of balance calibration checks.
(l) Calibration weights.Class S, standard calibration
weights to verify balance accuracy within weight range to be
used for weighing empty cream-test bottles and cream-test bottles with samples.

D. Determination
(a) Sample preparation and temperature adjustment.
Place sample in H2O bath at 38 1C. Level of H2O should
be at or above cream level. Mix sample 10 by inversion. If fat
line remains on inside surface of container, run hot tap H2O (ca
5060C) over outside surface for 1520 s. Mix sample thoroughly by inversion and weigh immediately. Do not allow samples to remain in H2O bath more than 15 min after reaching
38C.
Weigh 9 0.03 g tempered cream sample directly into 9 g
cream-test bottle, or 18 0.03 g into 18 g cream-test bottle.
Adjust temperature of cream in bottle to 2022C. Adjust temperature of H2SO4 to 2022C. Weigh some additional tempered cream for use in measurement of creamacid reaction
temperature.
(b) Measurement of creamacid reaction temperature and
determination of amount of H2SO4 to use.Before testing
group of samples, determine correct amount of acid to be used
in analysis by measuring creamacid reaction temperature.
Add 9 mL H2SO4 to 9 g bottle or 16 mL H2SO4 to 18 g bottle
from (a) containing tempered cream sample prepared for measurement of cream-acid reaction temperature. Add acid in one
delivery that washes all traces of cream into bulb and cleanly
layers acid under cream. Fully insert digital thermometer probe
down bottle neck and immediately shake by hand rotation until
all traces of curd disappear. Peak reaction temperature should
be within 93103C for 3045% cream (the lower the fat content, the higher the reaction temperature). Adjust amount of
H2SO4 added, until reaction temperature is within this range
and fat column is translucent golden-yellow to amber (the
lower the fat content, the lighter the color). Amount of acid
required may vary between technicians and between batches of
acid.
(c) Testing cream samples.To test bottle from (a) containing tempered cream, add appropriate amount of H2SO4 [as
determined in (b)] by delivering acid in one delivery that
washes all traces of cream into bulb and cleanly layers acid
under cream. Immediately shake bottle by hand rotation as in
(b), until all traces of curd disappear. Place bottle in Babcock
bottle shaker set at medium speed. Continue to add acid to all
samples, and then shake full set 1 additional min. Add 6 mL
distilled H2O at 56.561C [tempered in H2O bath either C(e)
or C(g)] to all bottles and then shake full set 1 additional min.
Place bottles in heated centrifuge, counter-balance, and centrifuge 5 min after proper speed is reached. Add distilled H2O

at 5961C until bulb of bottle is filled. Centrifuge 2 min. Add

distilled H2O at 5961C until top of fat column is between
4550% marks on bottle neck. Centrifuge 1 min longer at ca
60C. Transfer bottle to H2O bath maintained at 57.5 1C,
immerse bottle to level slightly above top of fat column, and
leave in H2O bath until column is at equilibrium and lower fat
surface assumes final flat form (5 min).
Remove 1 bottle from bath, wipe dry, and carefully add
room temperature red mineral oil to surface of fat column.
Add only a few drops of mineral oil into bottle just before reading, letting it flow down inside of neck to gently layer onto the
surface of the fat column. For purpose of measurement, surface
separating mineral oil and fat is regarded as representing upper
limit of column.
With aid of reading light and magnification, C(j), use dividers or calipers to quickly measure fat column (before it begins
to cool and contract). Place caliper points in vertical line on
neck of bottle, with 1 point at lowest surface of lower meniscus
and other point at interface between mineral oil and fat. Without
changing distance between 2 points on calipers, move calipers
down bottle neck until lower point rests in etched horizontal
graduation mark at 0%. Place upper point of calipers against
bottle graduation and read test in % to nearest 0.25%. Repeat
for each bottle.
Fat column, at time of measurement, should be translucent,
golden-yellow or amber, and free of visible suspended particles. Reject all tests in which fat column is milky or shows
presence of curd or charred matter, or in which meniscus is
indistinct or distorted; repeat test, adjusting volume of H2SO4
added, to obtain proper color and creamacid reaction temperature.
Ref.: J. AOAC Int. 79, 907(1996)
995.19 Fat in Cream, Mojonnier Ether Extraction
Method, IDFISOAOAC Method
First Action 1995
(Applicable to raw, homogenized, and heat-treated cream
containing 45% fat.)
(Caution: See Appendix B, safety notes on distillation, ammonium hydroxide, flammable solvents, diethyl ether, ethanol,
peroxides, and petroleum ether. Dispose of waste solvents according to applicable environmental rules and regulations.)
Method Performance:
sr = 0.125; sR = 0.151; RSDr = 0.330%; RSDR = 0.398%; r
= 0.354; R = 0.427.

A. Principle
Fat is separated by liquidliquid extraction from known
weight of cream. Ether extract is decanted into dry weighing
dish, and ether is evaporated. Extracted fat is dried to constant
weight. Result is expressed as % fat by weight.

B. Apparatus
(a) Flask.Mojonnier-style ether extraction flask with
volume of 2123 mL in lower bulb plus neck at bottom of flask.

Flask should have smooth, round opening at top that will seal
when closed with cork.
(b) Weighing dishes.Metal, 8.59.5 cm diameter and
4.55.5 cm tall; or 250 mL glass beakers.
(c) Calibration weights.Class S, standard calibration
weights to verify balance accuracy within weight range of empty
flasks, flask with sample, empty dishes, and dishes with fat.
(d) Analytical balance.To read to nearest 0.0001 g. Accuracy on verification, 0.0002 g. Check weekly and whenever balance is moved or cleaned. Keep record of balance calibration checks.
(e) Desiccator.For cooling weighing dishes to room
temperature after preliminary and final drying. Use coarse desiccant (mesh size 616) that contains minimum of fine particles
and that changes color when moisture is absorbed.
(f) Tongs.For handling weighing dishes.
(g) Hot plate.Steam bath or other heating device; for
evaporation of ether at 100C. Perform evaporation in hood.
(h) Corks.High quality natural cork stoppers (size 5) for
flasks. Soak corks in H2O several hours to improve seal.
(i) Vacuum or forced air oven.Vacuum oven capable of
maintaining temperature of 7075C at 50.8 cm (20 in.) of
vacuum, or forced air oven capable of maintaining temperature
of 100 1C.
(j) Water bath for tempering cream samples prior to weighing.With thermometer and device to maintain cream temperature of 38 1C.

C. Reagents
(a) Ethyl ether.ACS grade, peroxide free. No residue on
evaporation.
(b) Petroleum ether.ACS grade, boiling range 3060C.
No residue on evaporation.
(c) Ammonium hydroxide.Concentrated, ACS grade,
specific gravity 0.9.
(d) Ethyl alcohol.95%. No residue on evaporation.
(e) Distilled H2O.Free of oil and mineral residue.
(f) Phenolphthalein indicator.0.5% (w/v) in ethyl alcohol.

D. Determination
(a) Sample preparation.Place sample in H2O bath at 38
1C. Level of H2O should be at or above cream level. Mix
sample 10 by inversion. If fat line remains on inside surface
of container, run hot tap H2O (ca 5060C) over outside surface for 1520 s. Mix sample thoroughly by inversion and
weigh aliquot immediately. Do not allow samples to remain in
H2O bath more than 15 min after reaching 38C.
Weigh empty flask with clean dry cork stopper. Remove
stopper. Pipet into flask enough cream to yield 0.30.6 g extracted fat (e.g., ca 1 g 40% cream, 2 g 20% cream). Place stopper in flask and weigh to the nearest 0.1 mg. Check balance
zero between samples. Dilute test sample to ca 10 mL with distilled H2O at 2022C.
(b) Weighing dish preparation.Number clean weighing
dishes and pre-dry under same conditions used for final drying
after fat extraction. Be sure that all surfaces where weighing
dishes will be placed (i.e., hot plate, desiccator, etc.) are clean

and free of particulates. At end of oven drying, place pans in

room temperature desiccator and cool to room temperature. On
day of fat extraction, (c), weigh dishes to nearest 0.1 mg and
record weights. Check balance zero after weighing each pan.
Protect weighed pans from contamination with extraneous
matter.
(c) Fat extraction.To sample in flask add 1.5 mL NH4OH
and mix thoroughly. NH4OH neutralizes any acid present and
dissolves casein. Add 3 drops of phenolphthalein indicator to
help sharpen visual appearance of interface between ether and
aqueous layers during extraction. Add 10 mL ethyl alcohol,
stopper with H2O-soaked cork, and shake flask vigorously 15 s.
For first extraction, add 25 mL ethyl ether, stopper with
cork, and shake flask very vigorously 1 min, releasing built-up
pressure by loosening stopper as necessary. During vigorous
shaking, hold main body of flask horizontally with lower bulb
and stopper up. Add 25 mL petroleum ether, stopper with cork,
and repeat vigorous shaking for 1 min. Centrifuge flasks 30 s
at ca 600 rpm to obtain clean separation of aqueous (bright
pink) and ether phases. Decant ether solution into suitable
weighing dish prepared as in (b). When ether solution is decanted into dishes, be careful not to pour over any suspended solids or aqueous phase into weighing dish. Ether can be evaporated
at 100C from dishes while conducting second extraction.
For second extraction, add 5 mL ethyl alcohol, stopper with
cork, and shake vigorously 15 s. Next, add 15 mL ethyl ether,
replace cork, and shake flask vigorously 1 min. Add 15 mL petroleum ether, stopper with cork, and repeat vigorous shaking
for 1 min. Centrifuge flasks 30 s at ca 600 rpm to obtain clean
separation of aqueous (bright pink) and ether phases. If interface is below neck of flask, add H2O to bring level ca half way
up neck. Add H2O slowly down inside surface of flask so that
there is minimum disturbance of separation. Decant ether solution for second extraction into same weighing dish used for first
extraction.
For third extraction, omit addition of ethyl alcohol and repeat procedure used for second extraction. Completely evaporate solvents in hood on hot plate at 100C (avoid spattering).
Dry extracted fat plus weighing dish to constant weight in
forced air oven at 100 1C (30 min) or in vacuum oven at
7075C at >50.8 cm (20 in.) of vacuum for 7 min. Remove
weighing dishes from oven and place in desiccator to cool to room
temperature. Record weight of each weighing dish with fat.
Run 2 reagent blanks each day tests are conducted. To run
reagent blank, substitute cream with 10 mL H2O and run test as
described. Record weight of any dry residue collected and use
value in calculation. Reagent blank should be <0.0020 g residue. If reagent blanks for set of samples are negative, use negative number in calculation. (Note: To subtract negative number
[average weight blank residue] in equation below, add it to
[(weight dish + fat) weight dish].) Negative blank usually
indicates that dishes were not completely dry at start of determination or that balance calibration shifted between weighing
of empty pans and pans with fat. Cause of negative blanks and
excessively high positive blanks should be identified and corrected.

E. Calculation
Calculate fat content (%) in cream as follows:
Fat, % = 100

(wd + f) (wd) wb
wc

where wd = weight dish, f = fat, wb = average weight blank

residue, and wc = weight cream.
Ref.: J. AOAC Int. 79, 907(1996)
Results and Discussion
Data for the determination of fat content of cream by the
Babcock and ether extraction methods are presented in Tables 1 and 2, respectively. Invalid data and statistical outliers
are identified in these tables. The total amount of invalid and
outlier data were within the acceptable limits for collaborative
studies (9).
The statistical parameters of method performance for each
sample material are presented in Tables 3 and 4. Linear regression indicated no significant (P > 0.01) influence of fat concentration on any of the statistical parameters. Overall method performance was calculated from the square root of the average of
the squared deviations. Method performance for the milk Babcock
and ether extraction methods from a previous study (5) is indicated
in the tables for comparison with performance on cream.
On an absolute fat basis, the performance of the milk methods was about 7- to 10-fold better than the corresponding cream
methods. This was expected, as the repeatability and reproducibility of the methods is dependent on the sample fat content, the
test sample size (especially in the case of ether extraction
method), and resolution of the glassware (this affects only the
Babcock test). However, the performance of both cream methods was similar to, or better than, that of the corresponding milk
methods on a relative fat basis (RSDr and RSDR), indicating
acceptable method performance relative to the fat content of the
samples.
As in the milk methods (5), the performance characteristics
of the ether extraction method for cream were superior to those
of the cream Babcock method. Figure 1 illustrates that the
RSDr and RSDR values for both the milk and cream ether extraction methods are about half that of their corresponding Babcock methods. The ether extraction method is considered more
accurate, because it is a gravimetric determination. The Babcock method, on the other hand, is a volumetric determination
and uses glassware that estimates fat content in defined increments (0.05% fat for milk and 0.50% fat for cream). Additionally, the Babcock method requires certain assumptions such as
a single density for all milk fat.
A comparison of cream fat content estimation by the Babcock and ether extraction methods is presented in Table 5. Test
results using the Babcock method were consistently higher (average 0.277% fat) than those obtained by ether extraction. The
average difference was 0.730% (percentage of the fat content
of the samples) and was similar to the relative fat differences of
0.540% (5) and 0.782% (6) obtained between methods for testing milk.

The difference between the Babcock and ether extraction

tests is primarily caused by the volumetric scale on the neck of
the Babcock bottle, which was originally designed for milk
with an average fat density of 0.9000 g/mL at 60C (10). However, the density of the material extracted by the milk Babcock
test is closer to 0.8935 g/mL at 57.5C [calculated assuming
that the material in the Babcock column has an average density
at 60C of 0.8918 g/mL and coefficient of expansion of
75.58 105 mL/mL/C (10, 11)]. The results from both the
milk and cream collaborative study can be adjusted to reflect a
density of 0.8935 g/mL, using a correction factor of 0.9928 (derived from fat density, column volume, and sample size). When
this is done, the average difference between Babcock and ether
extraction methods is 0.007% fat for the 1988 milk collaborative study (5), 0.002% fat for the samples exchanges during
1989 through 1992 (6), and 0.002% fat for the present cream
collaborative study. Thus, adjusting the Babcock test to closer
reflect the true density of the material in the fat column seems
to be a satisfactory and justifiable means for bringing the Babcock test into closer agreement with ether extraction.
A number of different approaches to doing this are possible.
One approach would be to redesign the graduated scale on the
neck of the Babcock bottle. However, it would be impractical
and extremely costly to expect all laboratories that conduct the
Babcock test to purchase new glassware. A correction factor
could be applied to all Babcock tests to adjust for density, but
this approach is considered less than ideal for regulatory reasons. A third approach is to modify the temperatures at which
the Babcock test is conducted and the Babcock bottles are tempered prior to reading the column. Density is a function of temperature, and at 48C the material in the Babcock column
should be close to the density assumed by the glassware
(0.9000 g/mL). This last approach is currently being investigated, and will be followed up with a collaborative study if
initial investigations prove promising.
Recommendation
On the basis of the results of the collaborative study, the
Associate Referee recommends that the replacement method
description for determination of fat in cream by the Babcock
method be adopted first action as an improved method to replace 920.111BC and that the new method for determination
of fat in cream by the Mojonnier ether extraction method be
adopted first action.
Acknowledgments
We thank the Cornell University laboratory for their help
with the preliminary work, and the following U.S. Department
of Agriculture Federal Milk Markets and their associates, for
their cooperation in the sampling and/or laboratory analysis in
this study:
USDA Tulsa Milk Market Administrator Laboratory, Tulsa, OK
USDA Springfield Milk Market Administrator Laboratory,
Springfield, MO

USDA Kansas City Milk Market Administrator Laboratory,

Lenexa, KS
USDA Chicago Milk Market Administrator Laboratory,
Lisle, IL
USDA Cleveland Milk Market Administrator Laboratory,
Middleburg Heights, OH
USDA Atlanta Milk Market Administrator Laboratory, Norcross, GA
USDA Dallas Milk Market Administrator Laboratory, Carrollton, TX
USDA Phoenix Milk Market Administrator Laboratory,
Phoenix, AZ
USDA St. Louis Milk Market Administrator Laboratory, St.
Louis, MO
Barbano Laboratory, Cornell University, Food Science Department, Ithaca, NY
References
(1) Official Methods of Analysis (1990) 15th Ed., AOAC, Arlington, VA, sec. 920.111BC

(2) Official Methods of Analysis (1990) 15th Ed., AOAC, Arlington, VA, sec. 989.05
(3) International IDF Standard (1987) Cream: Determination of
Fat Content (Rose-Gottleib Reference Method), Brussels,
Belgium, 16C:1987, 17
(4) Official Methods of Analysis (1990) 15th Ed., AOAC, Arlington, VA, sec. 989.04
(5) Barbano, D.M., Clark, J.L., & Dunham, C.E. (1988) J. Assoc. Off. Anal. Chem. 71, 898914
(6) Lynch, J.M., Barbano, D.M., Healy, P.A., & Fleming, J.R.
(1994) J. AOAC Int. 77, 976981
(7) Lynch, J.M., Barbano, D.M., Houghton, G., & Fleming, J.R.
(1995) J. AOAC Int. 78, 463471
(8) Official Methods of Analysis (1990) 15th Ed., AOAC, Arlington, VA, sec. 920.111A
(9) AOAC INTERNATIONAL (1993) AOAC Official Methods
Program Manual on Development, Study, Review, and Approval Process for AOAC Official Methods, AOAC,
Arlington, VA
(10) Jenness, R., Herreid, E.O., Caulfield, W.J., Burgwald, L.H.,
Jack, E.L., & Tuckey, S.L. (1942) J. Dairy Sci. 25, 949960
(11) Jenness, R., & Herreid, E.O. (1945) J. Dairy Sci. 28, 591595

Table 1. Percentage of fat in cream determined by Babcock method

Laboratory
Material
1
2
3
4
5
6
7
8
9

29.75
30.00
32.00
32.00
34.00
34.00
35.75
36.00
39.00
38.75
41.50
41.50
41.50
42.00
a
48.50a
a
48.50a
45.50
45.50

30.00
30.00
32.25
32.00
34.50
34.50
36.00
35.75
39.00
39.00
a
42.50a
a
43.00a
42.00
42.00
44.00
44.00
45.50
45.75

30.00
30.50
32.00
32.00
34.50
34.50
36.00
36.00
39.00
39.00
41.00
41.50
42.00
42.00
44.00
44.00
45.50
45.50

30.00
30.00
32.00
32.50
34.50
34.50
36.00
36.00
38.75
39.00
41.00
41.00
42.00
42.00
44.00
43.50
45.00
45.50

30.00
30.50
32.25
32.50
34.00
34.50
36.00
35.50
39.00
38.75
41.75
41.50
41.75
42.50
42.50
43.00
45.75
46.25

29.50
30.00
32.00
32.00
34.00
34.25
35.75
36.00
38.75
38.75
41.50
41.50
41.50
41.50
43.50
43.50
45.50
45.25

30.00
30.00
32.50
32.25
34.50
34.50
36.00
36.00
38.75
39.00
41.75
41.50
41.75
41.75
44.00
44.00
45.25
45.50

29.75
30.00
32.00
32.00
34.25
34.50
35.75
35.75
39.00
38.75
41.25
41.50
42.00
42.00
44.00
44.00
46.00
45.50

29.50
30.00
32.25
32.00
34.50
34.50
36.00
35.50
38.75
38.75
41.25
41.25
42.00
42.00
44.00
44.00
46.00
45.75

29.50
30.00
32.00
32.00
34.75
34.25
35.75
36.00
38.50
39.00
42.00
41.25
42.25
41.75
44.50
44.00
45.50
45.25

Statistical outlier (Grubbs test).

Table 2. Percentage of fat in cream determined by ether extraction method

Laboratory
Material
1
2
3
4
5
6
7
8
9

a
b
c

29.370
29.894
31.768
31.572
33.855
34.059
35.419
35.394
38.266
38.809
b
40.812b
b
40.742b
b
40.864b
b
41.225b
43.274
43.293
44.943
44.788

29.653
29.730
31.732
31.702
34.004
33.981
35.364
35.409
38.626
38.494
41.037
41.067
41.358b
41.469b
43.517
43.472
44.994
44.957

29.792
29.764
31.885
31.874
34.175
34.221
35.566
35.520
38.617
38.638
41.107
41.148
41.543
41.527
43.572
43.587
45.208
45.102

29.767
29.934
31.898
31.873
34.162
34.168
35.606
35.484
38.626
38.575
41.148
41.145
41.530
41.568
43.596
43.569
45.031
45.039

29.956
29.874
31.930
31.939
34.293
34.219
35.516
35.679
38.821
38.104
41.175
41.232
41.601
41.514
43.722
43.621
45.235
45.154

29.806
29.666
31.931
32.000
34.175
34.112
35.584
35.664
38.753
38.687
41.220
41.197
41.512
41.602
43.733
43.535
45.162
45.176

29.821
29.818
31.996
31.888
34.265
34.286
35.565
35.678
38.831
38.708
41.274
41.249
41.451
41.702
43.604
43.843
45.388
45.219

29.864
29.956
31.919
31.982
34.420
34.150
35.633
35.581
38.639
38.731
41.201
41.295
41.605
41.439
43.617
43.796
45.179
45.362

28.974
29.767
31.904
31.923
34.139
34.120
35.557
35.616
38.743
38.745
41.179
41.250
41.545
41.552
c
43.514c
c
44.879c
45.156
45.058

Invalid data (reagent contamination).

Statistical outlier (Grubbs test).
Statistical outlier (Cochran test).

Table 3. AOAC statistical parameters by sample material for perentage of fat in cream determined by Babcock
method

Material
1
2
3
4
5
6
6
7
8
8
9

Data

No. of labs

No. of
accepted
values

All
All
All
All
All
All
Outlier removed
All
All
Outlier removed
All

10
10
10
10
10
10
9
10
10
9
10

20
20
20
20
20
20
18
20
20
18
20

Mean, %

sr

RSDr, %

sR

RSDR, %

29.950
32.125
34.375
35.875
38.863
41.550
41.417
41.913
44.275
43.806
45.563

0.262
0.158
0.177
0.194
0.168
0.250
0.236
0.231
0.194
0.204
0.231

0.876
0.492
0.514
0.540
0.432
0.602
0.569
0.550
0.437
0.470
0.506

0.264
0.192
0.224
0.194
0.168
0.501
0.273
0.248
1.549
0.466
0.294

0.882
0.597
0.653
0.540
0.432
1.205
0.660
0.592
3.499
1.072
0.645

0.734
0.443
0.495
0.542
0.470
0.700
0.660
0.645
0.542
0.572
0.645

0.740
0.537
0.628
0.542
0.470
1.402
0.765
0.694
4.338
1.315
0.823

Meana

38.209

0.209

0.548

0.272

0.712

0.592

0.769

Milkb

3.907

0.029

0.742

0.039

1.014

0.081

0.111

Sample mean calculated by averaging. All other means calculated from the square root of the average of the squared deviations. Values only
include data from sample materials 6 and 8 with outliers removed.
Determined in 1988 collaborative study (5).

Table 4. AOAC statistical parameters by sample material for percentage of fat in cream determined by Mojonnier
ether extraction method

Material
1
2
3
4
5
6
6
7
7
8
8
9

Data

No. of labs

No. of
accepted
values

All
All
All
All
All
All
Outlier removed
All
Outlier removed
All
Outlier removed
All

9
9
9
9
9
9
8
9
7
9
8
9

18
18
18
18
18
18
16
18
14
18
16
18

Mean, %

sr

RSDr, %

sR

RSDR, %

29.745
31.873
34.156
35.546
38.634
41.138
41.183
41.478
41.549
43.652
43.584
45.119

0.233
0.058
0.084
0.063
0.218
0.038
0.036
0.118
0.088
0.334
0.094
0.080

0.782
0.183
0.246
0.178
0.565
0.092
0.088
0.284
0.211
0.764
0.216
0.178

0.236
0.114
0.131
0.100
0.218
0.153
0.074
0.189
0.088
0.339
0.158
0.152

0.794
0.358
0.385
0.282
0.565
0.371
0.180
0.455
0.211
0.777
0.362
0.336

0.652
0.163
0.236
0.177
0.611
0.106
0.101
0.330
0.246
0.934
0.264
0.224

0.661
0.319
0.368
0.281
0.611
0.427
0.207
0.528
0.246
0.950
0.442
0.425

Meana

37.932

0.125

0.330

0.151

0.398

0.354

0.427

Milkb

3.886

0.015

0.396

0.020

0.512

0.044

0.056

Sample mean calculated by averaging. All other means calculated from the square root of the average of the squared deviations. Values only
include data from sample materials 6, 7, and 8 with outliers removed.
Determined in 1988 collaborative study (5).

Table 5. Comparison of fat content (%

%) of cream by the Babcock and Mojonnier ether extraction methods
Fat, %
Babcock method

Mojonnier method

Difference in fat results, %

Difference in fat results as

% of Mojonnier test results

1
2
3
4
5
6
7
8
9

29.950
32.125
34.375
35.875
38.863
41.417
41.913
43.806
45.563

29.745
31.873
34.156
35.546
38.634
41.183
41.549
43.584
45.119

0.205
0.252
0.219
0.329
0.229
0.234
0.364
0.222
0.444

0.689
0.791
0.641
0.926
0.593
0.568
0.876
0.509
0.984

Mean

38.209

37.932

0.277

0.730

3.907
3.739

3.886
3.710

0.021
0.029

0.540
0.782

Sample

Milk 1988
b
Milk 198992
a
b

Determined in 1988 collaborative study (5).

Determined bimonthly during 19891992 (6).

Figure 1. RSDr and RSDR values of Babcock and ether

extraction methods for milk and cream. Black bars
represent the Babcock method and white bars represent
the ether extraction method.