Size Distribution of Fat Globules in Human Colostrum, Breast Milk, and Infant Formula

J. Dairy Sci. 88:1927–1940
? American Dairy Science Association, 2005.
Size Distribution of Fat Globules in Human Colostrum,
Breast Milk, and Infant Formula
M. C. Michalski,1 V. Briard,1 F. Michel,1 F. Tasson,2 and P. Poulain2
1INRA-UMR 1253, Science et Technologie du Lait et de l’Oeuf Agrocampus 65, rue de Saint-Brieuc,
35042 Rennes Cedex, France
2Maternite´ de l’Hoˆtel Dieu, CHU de Rennes-2, rue de l’Hoˆtel Dieu, 35000 Rennes, France
ABSTRACT
INTRODUCTION
Only a few results are available on the size of human
Fat is a major component of human milk and is
milk fat globules (MFG), despite its signi?cance re-
composed of >98% triacylglycerols (Mulder and Wals-
garding fat digestion in the infant, and no data are
tra, 1974; Hamosh et al., 1985; Jensen et al., 1990).
available at <24 h postpartum (PP). We measured the
The nonpolar nature of milk lipids prevents solubility
MFG size distribution in colostrum and transitional
in the aqueous phase, within the mammary secretory
human milk in comparison with fat globules of mature
cells before secretion, as well as in milk (Hamosh et
milk and infant formula. Colostrum and transitional
al., 1999). Thus, fat globules are formed throughout
milk samples from 18 mothers were collected regularly
the mammary epithelial cell, grow in size as they move
during 4 d PP and compared with mature milk samples
toward the apical cell membrane, and are extruded
of 17 different mothers and 4 infant formulas. The size
into the alveolar lumen (Jensen et al., 1990; Keenan
distribution was measured by laser light scattering.
and Dylewski, 1995; Keenan, 2001; Ollivier-Bousquet,
For further characterization, the ?-potential of some
2002). During the extrusion process, the globule is
mature MFG was measured by laser Doppler electro-
enveloped by portions of the cell membrane, which
phoresis. The MFG diameter decreased sigmoidally in
becomes the milk fat globule (MFG) membrane. The
the ?rst days. At <12 h PP, the mode diameter was
core of the MFG contains triacylglycerols (Figure 1),
8.9 ± 1.0 µm vs 2.8 ± 0.3 µm at 96 h PP. Thus, the
and the membrane has the general composition of bio-
surface area of MFG increased from 1.1 ± 0.3 to 5.4 ±
logical membranes, i.e., phospholipid, cholesterol, gly-
0.7 m2/g between colostrum and transitional milk. In
coproteins, enzymes, etc. (Anderson and Cawston,
mature milk, the MFG diameter was 4 µm on average
1975; McPherson and Kitchen, 1983; Christie, 1995;
and increased with advancing lactation, whereas the
Keenan and Dylewski, 1995; Hamosh et al., 1999). The
droplets in infant formula measured 0.4 µm. The ?-
physico-chemical properties [structure, size distribu-
potential of mature MFG was ?7.8 ± 0.1 mV. The fat
tion, electrokinetic potential (or so-called ?-potential),
globules are larger in early colostrum than in transi-
and thermal behavior] of bovine MFG have been
tional and mature human milk and in contrast with
widely studied (Walstra, 1969; Mulder and Walstra,
the small-sized fat droplets in infant formula. Human
1974; Christie, 1995; Walstra, 1995; Lopez et al., 2000;
MFG also have a low negative surface charge com-
Michalski et al., 2002d; Briard et al., 2003).
pared with bovine globules. These structural differ-
Many studies concern the chemical nature and bio-
ences can be of nutritional signi?cance for the infant.
chemical signi?cance of human milk lipids (Emmett
(Key words: human milk, milk fat globule, size distri-
and Rogers, 1997; Francois et al., 1998; Villalpando
bution, ?-potential)
and del Prado, 1999; Fidler and Koletzko, 2000; Fran-
cois et al., 2003; Schweigert et al., 2004). However,
Abbreviation key: d32 = volume-surface average di-
despite the great signi?cance of the physical structure
ameter, d43 = volumic average diameter, MFG = milk
of milk fat regarding lipase acitivity, cholesterol avail-
fat globule, PP = postpartum, SSA = speci?c surface
ability and lipid absorption in the gastrointestinal
area.
tract of neonates (Bitman et al., 1983; Armand et al.,
1996; Armand, 1998; Armand et al., 1999; Hamosh et
al., 1999; Lo¨nnerdal, 2003), only a few researchers
have studied the size distribution of human MFG
Received January 13, 2005.
(Whittlestone and Perrin, 1954; Ru¨egg and Blanc,
Accepted February 23, 2005.
Corresponding author: M. C. Michalski; e-mail: marie-caroline.
1981, 1982; Simonin et al., 1984). Since the latter stud-
michalski@rennes.inra.fr.
ies, in which a microscope or a particle counter were
1927

---

1928
MICHALSKI ET AL.
Figure 1. Typical structure of the native milk fat globule. Schemes are not to scale.
used, laser light scattering devices have been devel-
MATERIALS AND METHODS
oped (McCrae and Lepoetre, 1996); the newest ones
allow a better study of MFG size distribution from
Breast Milk Samples
submicronic to micrometric particles thanks to 2 dif-
Colostrum and breast milk samples from 18 volun-
ferent laser beams (Michalski et al., 2001). To our
teer mothers, who had delivered at term and signed
knowledge, no MFG size data are available to date
an informed agreement for participating in the study,
regarding colostrum during the ?rst 24 h after deliv-
were collected twice a day [once in the morning and
ery, although it is known that colostrum composition
once in the evening from delivery (colostrum) to the
varies greatly within the ?rst 4 d postpartum (PP)
end of the 4-d stay in the hospital (transitional milk)].
(Jensen et al., 1990). Also, the surface properties of
Consistent characteristics of the volunteer mothers
human MFG, such as their electrokinetic- or ?-poten-
are shown in Table 1. Thirteen mothers (71%) deliv-
tial, which accounts for the surface charge of the glob-
ered with epidural analgesia; one of them delivered
ules and can affect globule interactions with proteins
by caesarean. All samples were collected with midwife
and lipases, have not been yet characterized.
supervision by manual expression in microtubes with-
The aim of the present study was to measure the fat
out preservative. For ethical reasons, samples were
globule size distribution in colostrum and transitional
collected from the sucked breast at the end of nursing
milk from mothers having delivered term infants dur-
so that the fat milk could be analyzed (Aksit et al.,
ing the ?rst 4 d of lactation and to compare it with fat
2002). This method is physiologically consistent as the
globule size measured from 1) mature human milk
milk obtained by manual or mechanical expression
samples from other mothers corresponding to 1 to 37
may not have the same composition as milk consumed
mo of lactation and 2) infant formula for term infants.
during breastfeeding itself (Lucas et al., 1977; Aksit
Typical physico-chemical parameters of the fat glob-
et al., 2002). Some samples were also collected from
ules were characterized (average diameters, speci?c
the opposite breast before nursing to check that fat
surface area, and span and ?-potential and fat content
globule size did not vary throughout the single feeding.
for the mature milk of one mother) and compared with
The samples were stored at 4°C until they were ana-
literature results regarding human, cow, and homoge-
lyzed in the laboratory the day after collection. Prelim-
nized cow milk.
inary experiments in our laboratory had shown that
Journal of Dairy Science Vol. 88, No. 6, 2005

---

SIZE OF HUMAN MILK FAT GLOBULES
1929
Table 1. Characteristics of the volunteer mothers (n = 18) whose colostrum and milk were analyzed during
the ?rst 4 d postpartum. Means are expressed ± SEM.
Minimum
Maximum
Mean
Age, yr
24
41
30.9 ± 1.0
Weight at parturition, kg
58
84
70.8 ± 2.3
Weight gain during pregnancy, kg
8.5
24
11.7 ± 1.0
Child rank
1
5
1.9 ± 0.3
Total previous nursing duration, mo
0
22
3.5 ± 1.5
Term, wk
37
41
40.0 ± 0.3
overnight storage at 4°C did not affect the fat globule
Standard parameters were calculated by the soft-
size. The study was in accordance with the rules of
ware: the mode diameter (diameter at the peak maxi-
the ethics committee of the hospital. All women were
mum); the volumic average diameter [d43 = ?(vi × di)/
fed the same hospital diet. The main sources of dietary
?vi (where vi is the volume of globules in a size class
fat were vegetable oil, cheese, yogurt, meat, and ?sh.
of average diameter di)]; the volume-surface average
As a complementary study, we collected the mature
diameter [d32 = ?vi/?vi/di)]; the speci?c surface area
milk of 17 other volunteer mothers whose stage of
(SSA = 6 × ??1 × d ?1
32
, where ? is the milk fat density);
lactation was different, between 1.5 and 37 mo.
and the distribution [span = (d0.9 ? d0.1)/d0.5, where d0.9
In addition to breast milk samples, 4 commercial
is the diameter below which lie 90% of the globule
infant milk formulas were studied [Gallia 1er aˆge, Da-
volume and, respectively, 10% for d0.1 and 50% for d0.5].
none, Paris, France; Guigoz Evolier 1er aˆge, Guigoz 2e
The mean free distance between fat globules (dfree) was
aˆge, Nestle´, Paris, France; Novalac 2e aˆge, Novalac,
calculated from our results and from the literature as
Paris, France) after dispersion in water at 37°C at 14
dfree = 0.225 × d32 × [(0.74/F) ? 1], where F is the typical
to 16% according to the manufacturer’s instructions.
sample fat content from the literature. A detailed pre-
Fat droplets in these formulas were made of emulsi?ed
sentation of these calculations is given by Walstra et
vegetable oils (no lactic fat).
al. (1969) and Ru
¨ egg and Blanc (1981).
The laser Doppler electrophoresis apparatus and
Particle Size and ?-Potential Measurements
method for ?-potential analysis have been described
in details by Michalski et al. (2002d). Measurements
The apparatus and method for particle size analysis
were performed with a ZetaSizer 3000HS (Malvern).
have been described in detail by Michalski et al. (2001).
The fat globule size distribution was measured by laser
Biochemical Analyses
light scattering using a Mastersizer2000 (Malvern,
UK) with 2 laser sources, allowing the characterization
The fat content of mature milk samples was mea-
of micronic as well as submicronic populations. To pre-
sured by the Gerber standard method (FIL, 1997).
vent artefacts regarding submicronic globules, the ca-
sein micelles (usually ?150 nm) were dissociated by
Statistical Analyses
diluting the sample in 35 mM EDTA (pH 7) prior to
measurement. Dispersion in 0.1% SDS allowed dissoci-
Size distribution parameters among different times
ation of clusters in homogenized infant formulas (Mi-
PP, different breasts, or different milks were compared
chalski et al., 2001). For the software to calculate parti-
by using Student’s t-test (Rice et al., 2000). Statistical
cle size distributions from the measured laser light
signi?cance was determined at P < 0.05. The Systat
scattering patterns, it is necessary to know the refrac-
program (SSI, Richmond, CA) was used. Trend curves
tive index of the particles. In preliminary experiments,
for the evolution of fat globule diameter and SSA vs.
we found that the same refractive index, such as that
time PP were obtained using the TableCurve 2D pro-
of cow milk fat, could be used for breast MFG of women
gram (v. 4; SSI).
fed a Western diet rich in dairy products (i.e., 1.458
and 1.460 for milk fat at 633 and 466 nm, respectively).
RESULTS
Refractive index increased by only 0.002 for women
Fat Globule Size Distribution of Colostrum
fed a diet without dairy products and rich in ?sh and
and Transitional Milk
vegetable oils, which was not found to affect size distri-
butions signi?cantly. Because infant formulas contain
Figure 2 shows the typical evolution of the MFG
a mixture of palm, sun?ower, coconut, and rapeseed
size distribution of 2 mothers during early lactation.
oils, their mean refractive index of 1.462 was used.
Shortly after delivery, no fat globule <1 µm could be
Journal of Dairy Science Vol. 88, No. 6, 2005

---

1930
MICHALSKI ET AL.
Figure 2. Examples of milk fat globule size distributions by volume (left column = mother #13; right column = mother #18) during the
?rst lactation days.
Journal of Dairy Science Vol. 88, No. 6, 2005

---

SIZE OF HUMAN MILK FAT GLOBULES
1931
fat globules <1 µm during the ?rst lactation days can
be seen in Figure 3B.
Diameters and SSA of Fat Globules in Colostrum
and Transitional Milk
Table 2 shows no signi?cant difference of fat globule
size parameters between milk expressed at the end of
nursing on the sucked breast and milk expressed be-
fore nursing on the opposite breast. Additional experi-
ments in our laboratory did not reveal any signi?cant
difference in MFG size during one nursing for mature
milk (results not shown). On the whole, these ?ndings
are consistent with Whittlestone and Perrin (1954),
who found no difference in fat globule size at the begin-
ning and end of feeding for human milk by microscopic
observations. No clear-cut explanation is available to
date to explain this lack of difference.
Figure 4 shows some evolution curves of MFG mode,
d43, SSA, and span as a function of the early lactation
stage. (Seven typical examples were chosen for each
size distribution parameter, representative of the dif-
ferent curve shapes observed in the study.) Mode and
d43 both decreased with early lactation stage during
the ?rst 4 d in a nearly linear, exponential, or sigmoi-
dal manner, depending on the mother. The SSA of
MFG increased up to 9-fold from colostrum to transi-
tional milk, as a result of the decrease of the particle
size and increase of the volume fraction of globules <1
µm (Figure 3B). Most of the time, this increase was
Figure 3. Milk fat globule size distribution (mother #18). A) Distri-
bution by number (thick line = delivery day evening; thin line = fourth
sigmoidal and could correspond to the onset of lacta-
day morning). B) Cumulative distribution by volume (thick line =
tion. In most cases, the size distribution span in-
delivery day evening, dashed line = second day morning, and thin
creased from delivery to 4 d PP, although large discrep-
line = fourth day morning).
ancies were observed among mothers.
One of the mothers (mother #10) presented signi?-
cantly larger MFG than the 17 other mothers, al-
observed. The mode diameter of the main fat globule
though her characteristics were in the average (31 yr
population decreased with the lactation stage together
old, 63 kg at delivery with 10-kg weight gain during
with the appearance of populations of smaller MFG
pregnancy, third child, 8 previous mo of lactation, 39-
<1 µm with a mode diameter around 200 nm.
wk gestation time, delivery with epidural, no caesar-
We should highlight that the size distribution by
ean). Moreover, she gave fewer milk samples than
volume, measured by laser light scattering, does not
other mothers. Therefore, the size distribution data
provide the same information as the size distributions
corresponding to this mother were not taken into ac-
by number, provided with a particle counter. Figure
count further for means and statistical analyses.
3A shows the number-transformed size distribution
Figure 5 shows the averaged mode, d43, and SSA of
of MFG (calculated by the apparatus software) from
MFG as a function of the early lactation stage. Differ-
mother #18. Although most of the number of globules
ent trend curves were ?tted to the weighted averaged
were ?10 µm on the evening of the delivery day, most
data. Results were best ?tted with a sigmoidal de-
of them were numerous and <1 µm on the fourth day,
crease of diameters and a sigmoidal increase of SSA.
despite the <20% representation of the total fat volume
Average mode and SSA values at speci?c stages PP
(Figure 3B). This is explained by the fact that 1 globule
are reported in Table 3, together with the increase of
of 10-µm diameter represents the same volume as 103
SSA from early colostrum to 4-d milk. Average values
globules of 1-µm diameter, although being numerically
before 48 h PP were not signi?cantly different, as well
in the minority. The increased volumic percentage of
as after 60 h PP. However, the decrease of diameters
Journal of Dairy Science Vol. 88, No. 6, 2005

---

1932
MICHALSKI ET AL.
Table 2. Comparison between mode, volumic average diameter (d43), and speci?c surface area (SSA) of the
fat globule population measured from milk extracted at the end of nursing from the sucked breast or from
the opposite breast during the ?rst 4 d postpartum (PP). Means are expressed ± SEM.1
Mode
d43
SSA
Time PP
Sucked
Opposite
Sucked
Opposite
Sucked
Opposite
(µm)
(m2/g)
15 h
11.6 ± 1.1
8.8 ± 1.7
—
—
—
—
50 h
7.6 ± 2.4
7.8 ± 1.2
11.2 ± 2.4
13.3 ± 0.0
2.6 ± 1.5
1.6 ± 0.0
75 h
3.7 ± 0.4
3.7 ± 0.4
—
—
—
—
80 h
3.6 ± 0.5
3.2 ± 1.1
5.3 ± 0.9
5.9 ± 2.0
4.0 ± 0.9
6.4 ± 1.9
90 h
3.3 ± 0.3
2.5 ± 0.1
4.3 ± 0.6
3.9 ± 0.6
5.6 ± 0.9
4.6 ± 1.7
1n = 3 to 8 depending on time PP. P > 0.05 between each sucked and opposite breast.
and increase of SSA were both signi?cant between 48
the absolute value (negative surface charge) of which
and 72 h PP, which is consistent with the sigmoidal
is lower than the mean value reported for cow MFG
correlations observed previously (Figure 5). This dif-
of ?13.5 mV (Michalski et al., 2002d). The absolute
ference should correspond to the average appearance
value of ?-potential of human MFG was also lower
of the transitional milk after colostrum.
than the value for homogenized milk fat droplets of
?20 mV (Michalski et al., 2002d), such as for commer-
Fat Globule Size and ?-Potential in Mature
cial MFG.
Milk Throughout Lactation
Particle Size Distribution of Fat
Figure 6A shows that the mode diameter of mature
Droplets in Infant Formula
MFG from a single mother increased signi?cantly from
3 to 20 mo lactation. Variations in diameter among
Fat droplets in infant formula are not native MFG
milks extracted in the morning, midday, and evening
but homogenized ones, produced by emulsi?cation of
were not found to be signi?cant. The range of fat glob-
vegetable oils using a homogenizer. The resulting par-
ule diameter for this mother is somewhat higher than
ticles can be coated with phospholipids or proteins,
reported in the literature (Ru¨egg and Blanc, 1981).
depending on the formulation. Fat droplets in infant
Plotting all together the diameter and SSA of mature
formula were much smaller than in human colostrum
MFG from different mothers at different stages of lac-
and milk (Table 4). The fat droplet size distribution
tation (Figure 6B), the mode tended to increase, and
from 2 different brands is presented in Figure 7. They
SSA tended to decrease, between 2 and 12 mo of lacta-
were mainly constituted of submicronic droplets,
tion. No differences were observed after 12 mo because
which is consistent with previous studies on homoge-
of the smaller number of samples collected. In France,
nized milk using the same measurement apparatus
only 48.5% of mothers breastfeed after delivery, and
(Michalski et al., 2002c).
only a few nurse >3 mo (Beaufre`re et al., 2000).
The fat globule size was found to increase with ma-
DISCUSSION
ture milk fat content for one single mother who had
Parameters of the Fat Globule Size
given entire milk samples collected by mechanical ex-
Distribution in Different Milks
pression (Figure 6C). This relationship between milk
fat content and globule diameter is also known for cow
The present results of human fat globule size within
milk (Wiking et al., 2004). This means that when milk
the ?rst 48 h PP are new compared with the literature
fat content increases, the volume of fat globules in-
(Ru
¨ egg and Blanc, 1981). Our results regarding colos-
creases rather than their number. This diameter de-
trum between 2 and 5 d PP and average mature milk
pendency on fat content can also explain some of the
(Table 5) are consistent with those of Ru¨egg and Blanc
observed variations of mode between morning and
(1981) and Simonin et al. (1984). Fat globule sizes
midday milk (Figure 6A) and the discrepancies be-
presented in the literature as “initial” (Simonin et al.,
tween various mothers at the same stage of lactation
1984) were smaller, and the SSA was larger than in our
(Figure 6B).
study. However, this difference is likely to be caused by
Figure 6A shows that the ?-potential of mature MFG
the different early times PP between the 2 studies and
did not increase signi?cantly from 8 to 15 mo of lacta-
the small number of initial samples in the study by
tion. The average ?-potential value was ?7.8 ± 0.1 mV,
Simonin et al. (1984). Moreover, some discrepancies
Journal of Dairy Science Vol. 88, No. 6, 2005

---

SIZE OF HUMAN MILK FAT GLOBULES
1933
Figure 4. Mode diameter, volumic average diameter (d43), speci?c surface area (SSA), and span of the milk fat globules of individual
mothers during the ?rst lactation days. Mother #2 (--?--), mother #3 (-?-), mother #6 (..?..), mother #10 (..×..; on secondary axis), mother
#13 (–?–), mother #15 (–?–), mother #16 (..?..), mother #18 (–?–), mother #21 (-?-), and mother #25 (–?–).
Journal of Dairy Science Vol. 88, No. 6, 2005

---

1934
MICHALSKI ET AL.
Figure 5. Average mode, average volumic average diameter (d43), and average speci?c surface area (SSA) during the ?rst lactation days
calculated from the data of 17 mothers. Exponential trend curve (y = a × e?b × x) calculated with unweighted data: R2mode = 0.90, R2d43 =
0.82; R2S = 0.60 (?). Exponential trend curve (y = a + b × e?c × x) calculated with weighted data: R2mode = 0.90, R2d43 = 0.97; R2S = 0.84
(- - -). Sigmoidal trend curve [y = a + b × (1 + e?(x ? c)/d)?1] calculated with weighted data: R2mode = 0.92, R2d43 = 0.99; R2S = 0.90 (
).
Journal of Dairy Science Vol. 88, No. 6, 2005

---

SIZE OF HUMAN MILK FAT GLOBULES
1935
Table 3. Characteristic parameters of the milk fat globule population at various stages during the ?rst 4
d postpartum (PP). Mode = modal diameter of the main milk fat globule population; SSA = speci?c surface
area; RI-S = relative increase of SSA; and I-S = increase of SSA. Means are expressed ± SEM.1
<12 h PP1
36 to 48 h PP
60 to 72 h PP
>96 h PP
Mode, µm
8.9a ± 1.0
7.7a ± 0.7
4.4b ± 0.4
2.8c ± 0.3
SSA, m2/g
1.1a,2 ± 0.3
2.6a,2 ± 0.7
4.5b ± 0.7
5.4b ± 0.7
RI-S, %/h
4.0 ± 1.0
I-S, m2/g per h
5.3?10?2 ± 0.7?10?2
a,bValues in a row with different superscripts are different (P < 0.02).
1n = 7 to 16 depending on time PP.
2P = 0.06.
can be explained by the different techniques used to
shown before between 48 h (d43 = 7.8 ± 1.9 µm) and
measure particle size distributions. In the studies by
60 h (d43 = 4.3 ± 0.6 µm). This ?nal value is consistent
Ru
¨ egg and Blanc (1981), a particle counter was used,
with Ru
¨ egg and Blanc (1981) as well as size distribu-
instead of a 2-laser light scattering apparatus in the
tion parameters of mature human milk (Table 5). In
present work. Therefore, raw data in the literature
short, we have shown that the human MFG size ?rst
(Ru
¨ egg and Blanc, 1981) were obtained as size distri-
decreased during the ?rst 4 d of lactation, and that it
butions by number, and the volumic distributions pre-
subsequently increased in mature milk from the ?rst
sented were obtained after mathematical transforma-
month until the end of lactation.
tion of the former. Moreover, Ru¨egg and Blanc (1981)
The order of magnitude of human MFG diameter is
suggested the predominance of the smallest fat glob-
similar to that of bovine MFG (Table 5). However,
ules (<1 µm) in colostrum. However, it was not clear
the diameter of mature human MFG increases with
in their methods whether casein micelles were dissoci-
advancing lactation, conversely to bovine MFG
ated prior to particle counting. If not, because milk
(Mulder and Walstra, 1974; Ru¨egg and Blanc, 1981).
fat content is lower in early colostrum, some of the
Literature supports the theory that the secretory cells
submicronic particles that were observed in their re-
of the mammary gland are not able to increase the
sults might well be casein micelles. Finally, in the
production of membrane material when the fat yield
study of Ru
¨ egg and Blanc (1981), submicronic popula-
increases, which occurs with advancing lactation. In-
tions were estimated by extrapolation instead of being
stead, the fat droplets have to grow larger before they
truly measured such as in our study.
are covered with plasma membrane in the secretory
In the present study, the appearance of subpopula-
apical membrane (Wiking et al., 2004), which would
tions of different sized fat globules (Figure 3B) is con-
explain the correlation between mature milk fat con-
sistent with Ru
¨ egg and Blanc (1981). However, in the
tent and MFG size (Figure 6C). The release of lipid
latter reference, the number of small globules de-
globules and the ?nal size of the MFG may also be
creases signi?cantly during the early stages of lacta-
affected by hormones, such as prolactin and oxytocin
tion. Using a ?ner measurement technique with disso-
(Ollivier-Bousquet, 2002). It is possible that the initial
ciation of casein micelles and at earlier stages of lacta-
tion, we show a converse trend: the smallest MFG
decrease in fat globule diameter during the ?rst lacta-
appear progressively from delivery to 4 d PP. Yet, our
tion days is due to the maturation of the mammary
results agree with Ru
¨ egg and Blanc (1981) regarding
gland that progressively provides more and more
the decrease in frequency of the largest particles dur-
membrane to cover MFG, while the subsequent in-
ing early stages of lactation. The higher amount of
crease in mature fat globule size during later months
large fat globules in early colostrum can be accounted
of lactation is due to the increased fat content and
for if we assume that the immature mammary gland
limited available membrane coverage. Moreover, in
provides the fat droplets with incomplete membrane
cow milk, fatty acid composition is linked to breed,
coating so that coalescence of small globules occurs
feed, fat yield, and fat globule size (Briard et al., 2003;
(Simonin et al., 1984).
Mulder and Walstra, 1974; Wiking et al., 2004). It
Averaging particle size results between 2 and 5 d
would be interesting to study the relationship between
PP (Table 5), we obtain larger average d
the fat content and fat globule size distribution of hu-
43 values than
Ru
¨ egg and Blanc (1981). As mentioned previously, this
man milk and its fatty acid composition with mothers
could be due to the different techniques used. More-
under a controlled diet. The effect of the duration of
over, this time range corresponded to the sigmoidal
previous lactations on MFG size should also be
decrease that we highlight and that had never been
studied.
Journal of Dairy Science Vol. 88, No. 6, 2005

---

1936
MICHALSKI ET AL.
Figure 6. A) Mode diameter of mature milk fat globules (>3 mo postpartum) of one single mother during lactation, from morning milk
(?), midday milk (?), and evening milk (?) and ?-potential of mature milk fat globules from midday milk (?). Exponential trend curve (y =
a × e?b × x) for midday milk mode (R2 = 0.77). B) Mode diameter (?) and speci?c surface area (SSA; ?) of mature milk fat globules (1.5 to
37 mo postpartum) from 16 different mothers at various stages of lactation (lines are to guide the eye). C) Mode diameter of mature milk
fat globules as a function of milk fat content measured from one single mother throughout lactation.
Journal of Dairy Science Vol. 88, No. 6, 2005

---