International Journal of Food Microbiology 85 (2003) 137 149

www.elsevier.com/locate/ijfoodmicro

Microbiology of wheat and flour milling in Australia

Lana K. Berghofer a, Ailsa D. Hocking a,*, Di Miskelly b,c, Edward Janssona,1
a

Food Science Australia, P.O. Box 52, North Ryde, NSW 1670, Australia
b
Allied Mills, P.O. Box 1, Summer Hill, NSW 2130, Australia
c
Quality Wheat CRC, Locked Bag No. 1345, North Ryde, NSW 1670, Australia
Received 19 April 2002; received in revised form 30 October 2002; accepted 10 November 2002

Abstract
A survey was undertaken to determine the microbiological status of Australian wheat and the distribution of microorganisms
in the flour milling fractions and end products. A total of 650 milling process and end product samples was obtained from nine
flour mills located in New South Wales (4), Queensland (2), Victoria (2) and Western Australia (1) during the 1997 1998 and
1998 1999 wheat seasons. Most frequent (modal) counts in wheat and flour were, respectively, as follows: aerobic mesophilic
plate count, 105 and 102 colony forming units/gram (cfu/g); coliforms, 10 and 1 most probable number/gram (MPN/g); Bacillus
spp., 104 and 102 cfu/g; B. cereus, 1 and 0.1 MPN/g; mesophilic aerobic spores, 10 and 1 cfu/g; aerobic thermophiles, both 10
cfu/g; yeasts, 103 and 102 cfu/g, and moulds, 103 and 102 cfu/g. Bacillus spp., coliforms, yeasts and moulds were the most
frequently detected microorganisms throughout the survey. The most common moulds isolated were Aspergillus, Penicillium,
Cladosporium and Eurotium spp. Environmental serovars of Salmonella were isolated from two samples. Escherichia coli and
B. cereus were present at very low levels, a majority of positive samples being at the minimum level of detection (3 and 0.3
MPN/g, respectively). As wheat grain layers are separated, surface-adhering contaminants are concentrated in end product bran,
wheat germ and pollard, which comprise the outer layers of the grain. Consequently, the inner endosperm fraction contains
lower microbial counts, and flour is the cleanest end product of the milling process. Higher microbiological counts midstream in
the milling process indicate that equipment contamination may contribute to microbiological contamination; however, the
microbiological quality of incoming wheat has a strong influence on the ultimate quality of milling end products.
D 2002 Elsevier Science B.V. All rights reserved.
Keywords: Wheat; Flour; Flour milling; Bran; Wheat germ; Coliforms; Escherichia coli; Bacillus cereus; Bacillus spp.; Yeasts; Moulds;
Salmonella

1. Introduction
Flour is generally regarded as a microbiologically
safe product as it is a low water activity commodity.
* Corresponding author. Fax: +61-2-9490-8581.
E-mail address: ailsa.hocking@csiro.au (A.D. Hocking).
1
Current address: SafeFood NSW, 179 Elizabeth Street, Sydney,
NSW 2000, Australia.

Although the growth of pathogenic bacteria may not

be supported under such conditions, pathogens that
contaminate flour may survive for extended periods.
There are few reported incidents of food poisoning
resulting from contaminated flour. Australian, European and US studies indicate that Salmonella spp.,
Escherichia coli, Bacillus cereus and spoilage microorganisms are present in wheat and flour at low levels
(Cicognani et al., 1975; Ottogalli and Galli, 1979;

0168-1605/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0168-1605(02)00507-X

138

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

Spicher, 1986; Eyles et al., 1989; Richter et al., 1993).

In 1952, an outbreak of Salmonella Paratyphi B phage
type 1 occurred in New South Wales, Australia, where
flour was implicated, but the organism was never
isolated from the suspected flour (Anonymous, 1958;
Dack, 1961).
Flour milling begins with cleaning and scouring of
wheat grains to separate and remove non-wheat material. Australian wheat is stored under dry conditions
(usually 8 12% moisture, aw between 0.40 and 0.65)
to prevent development of fungal growth. Such dry
wheat is unsuitable for milling as it is too brittle. After
initial cleaning, before wheat enters the mill, water is
added, usually by means of a fine spray, to increase
the moisture content of the wheat to 14 15%, aw
0.68 070. The amount of water is carefully calculated
from the initial moisture content of the wheat (usually
measured by NIR spectrometry). This increases the
plasticity of the outer bran layer of the grain, preventing it from fracturing during milling and ensuring
easier separation from the endosperm (flour) later in
the milling process. To ensure even penetration of
water into the bran layer, conditioned wheat is held in
large conditioning bins, usually overnight, but sometimes for up to 24 36 h, depending on wheat type and
initial moisture content.
During milling, grains are broken open and undergo
a sequence of reduction, grinding and sifting operations to separate endosperm from outer grain layers.
Inner endosperm fractions are ground to produce
semolina, and then flour. Outer grain layers comprising bran, wheat germ and pollard are removed by
sifting. These operations create a considerable amount
of heat, so moisture condensation in the break rolls,
reduction rolls and sifters can sometimes lead to buildup of flour residues inside equipment. Microorganisms, particularly fungi, may become established in
these moister residues, causing contamination of the
mill products.
There is little information on the microflora of
Australian wheat and flour or the influence of milling
practices on the microbiological quality and safety of
flour and other end products. Substantial amounts of
data were generated in the USA by Hesseltine and
Graves (1966); however, the relevance of this study to
current Australian flour milling practices is limited.
There may also be a significant body of data from ongoing monitoring and quality assurance by the flour

milling industry that is not in the public domain. The

one Australian published study on the microbiological
status of Australian flour and the effects of milling
procedures tested only 24 samples (Eyles et al., 1989).
This study was undertaken to investigate the level,
nature and distribution of spoilage and pathogenic
microflora through certain mill fractions, and examine
the influence of various stages of the milling process
on the microbiological quality of end product flour.

2. Materials and methods

2.1. Commercial milling survey
Nine flour mills from three Australian milling
companies were sampled on a rotational basis,
approximately every 9 10 weeks. In each round
of testing, employees from each of the nine mills
collected between 8 and 14 samples throughout their
regular milling process. Sample points chosen by the
participating mills varied, but generally consisted of
the following: wheat prior to conditioning, wheat after
conditioning, grist after first break, grist after coarse
reduction, grist after fine reduction, semolina, end
product flour, bran, pollard and wheat germ. Samples
of 100 500 g were collected in plastic airtight bags.
A total of seven sampling rounds was completed.
The mill locations that were chosen were geographically diverse to encompass most of the major
wheat-growing areas. Four mills were located in New
South Wales, two in Queensland, two in Victoria and
one in Western Australia. The mills included those
using older milling equipment, and mills with recently
installed milling technology. The sampling period
commenced in November 1997 and concluded during
June 1999, spanning the 1997 1998 and 1998 1999
wheat harvests.
2.2. Microbiological testing
Samples were transported at ambient temperature
to Food Science Australia, North Ryde, NSW, within
48 h of collection. They were stored at 1 jC and
tested, usually within 7 days of receipt, for mesophilic
aerobic plate count, coliforms, Bacillus spp., B. cereus, mesophilic aerobic spores, thermophilic aerobic
plate count, yeasts, moulds, E. coli and Salmonella

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

spp. Testing for Salmonella spp. and thermophilic

aerobes ceased after round 6 as very low numbers
were consistently obtained.
All microbiological media used were purchased
from Oxoid Australia unless otherwise stated and
prepared as per suppliers instructions. All control
cultures were obtained from the CSIRO Food Science
culture collection (FRR), North Ryde, NSW, Australia.
An initial dilution was prepared from each sample
using 25-g sample and 225-ml 0.1% aqueous peptone
solution, pH 7.1 (Oxoid bacteriological peptone).
Samples were rehydrated at room temperature for 15
min prior to homogenising. Serial 10-fold dilutions
were then prepared using 0.1% peptone solution.
Aerobic mesophilic counts were performed following
Australian Standard 1766.2.1 (Standards Australia,
1991a). Dilutions and homogenates were tested with
plate count agar (PCA) using the pour plate technique
(Standards Australia, 1991b). Plates were incubated at
30 jC for 72 h and colonies counted. Thermophilic
aerobic viable counts were tested with PCA using the
pour plate technique. Dilutions and homogenates were
incubated aerobically at 55 jC for 72 h. Coliform and
E. coli testing was conducted in accordance with
Australian Standard 1766.2.3 three tube MPN method
(Standards Australia, 1992). Further information on
Australian Standard methods is available on the Standards Australia website (http://www.standards.com.au).
Bacillus spp. vegetative cells and spores were
enumerated by spread-plating homogenates and dilutions onto dextrose tryptone agar (DTA) following
Australian Standard 1766.1.4 (Standards Australia,
1991c). DTA plates were incubated aerobically at 37
jC for 48 h. Spore production was confirmed by
microscopic examination. Aerobic, Gram-positive,
catalase-positive, spore-forming rods were counted
as Bacillus spp. B. cereus enumeration followed the
Australian Standard 1766.2.6 MPN method (Standards Australia, 1991d) using double-strength tryptone
soya broth containing polymyxin B sulphate (Sigma,
St Louis, MO, USA) for enrichment of homogenates
and dilutions.
For mesophilic aerobic spore testing (adapted from
Stevenson and Segner, 1992), 10 ml of the initial 1:10
dilution was placed into a 250-ml flask containing 100
ml of DTA. Each flask was heat treated at 80 jC for
30 min, cooled and poured into five sterile Petri
dishes. Plates were incubated aerobically for 48 h

139

and colonies typical of mesophilic aerobic spore

formers counted.
Isolation of yeasts and moulds followed Australian
Standard 1766.2.2 (Standards Australia, 1997) using
dichloran rose-bengal chloramphenicol (DRBC) agar
and dichloran 18% glycerol (DG18) agar. Moulds
were identified to genus using the methods of Pitt
and Hocking (1997).
Examination for Salmonella spp. was in accordance
with Australian Standard 1766.2.5 (Standards Australia, 1991e), using 25-g samples. API 20E identification kits (bioMerieux, Marcy LEtoile, France) were
used to confirm presumptive colonies and identify
Salmonella to genus level. Presumptive isolates were
then sent to the Institute of Medical and Veterinary
Science in Adelaide, South Australia, for serotyping.
The water activity of all samples was measured
using a Novasina Humidat-RC analyser (Novasina,
Zurich, Switzerland) or Aqualab CX-3 dew point
instrument (Decagon, Pullman, Washington, USA).

3. Results and discussion

3.1. Microflora survey
Target microorganisms were detected in milling
and end products at varying levels. Table 1 shows a
comparison of the percentage of positive samples as a
proportion of the total number tested for each organism. Tables 2 and 3 illustrate the range and modal
(most frequently occurring) counts obtained for all
microorganisms at each stage of the milling process.
Table 4 provides a comparison of the quality of flour
milled from wheat with low and high mesophilic
aerobic counts.
Most products at all stages of the milling process
contained coliforms (70 98%), Bacillus spp. (70
94%) and B. cereus (64 94%, Table 1). Although
most samples contained a varied population (Tables 2
and 3), microbial levels were generally low, with
yeasts, moulds, coliforms and Bacillus spp. the most
frequently detected microbes. Mesophilic aerobic
spores and thermophilic aerobes were commonly
detected in wheat before and after conditioning and
in end product bran, but were low in end product flour
(modes = 1 and 10 cfu/g, respectively).

140

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

Table 1
Percentage of milling process and end product samples positive for the presence of the various classes of microorganisms tested
Organism

% Positive samples
Wheat

Coliforms
E. coli
Bacillus spp.
B. cereus
Mesophilic aerobic
spores
Aerobic
thermophiles
Yeasts
Moulds

After
conditioning

First break

Fine
reduction

Flour

Bran

Wheat germ

93
0
91
81
95

93
14
83
64
98

70
4
76
78
87

81
5
70
70
79

82
1
65
93
77

89
4
94
94
94

98
11
82
64
93

50

65

61

35

29

53

100
100

97
99

76
94

40
93

45
96

92
96

66
100

Table 2
Number of samples (N), range and mode (most frequently occurring log10 value) of microbiological counts detected in milling process samples
Organism
a,b

Aerobic mesophiles

Coliformsc,d

E. colic,d

Bacillus spp.a,b

B. cereusc,e

Mesophilic aerobic sporesa,f

Aerobic thermophilesa,b

Yeastsa,g

Mouldsa,g

a
b
c
d
e
f
g

N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode

Colony forming units/gram (cfu/g) sample.

Detection limit 10 cfu/g.
Most probable number/gram (MPN/g) sample.
Detection limit 3 MPN/g.
Detection limit 0.3 MPN/g.
Detection limit 1 cfu/g.
Detection limit 100 cfu/g.

Wheat

After conditioning

First break

Fine reduction

58
101 106
105
58
100 103
101
58
< 100
< 100
58
102 107
104
58
10 1 101
100
58
100 103
101
46
101 103
101
58
102 105
103
58
102 105
103

90
102 107
104
90
100 103
101
90
100 102
100
90
102 107
102
90
10 1 101
10 1
90
100 102
101
71
101 103
101
90
102 106
103
90
102 106
103

54
101 106
104
54
100 103
101
54
100 101
100
54
102 105
102
54
10 1 102
10 1
54
100 102
100
41
101 103
101
54
102 103
102
54
102 106
102

43
101 105
103
43
100 103
100
43
100
100
43
102 105
102
43
10 1 101
100
43
100 102
100
33
101 102
101
43
102 103
102
43
102 104
102

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

Table 3
Number of samples (N), range and mode (most frequently occurring
log10 value) of microbiological counts detected in end products
Organism
Aerobic
mesophilesa,b
Coliformsc,d

E. colic,d

Bacillus spp.a,b

B. cereusc,e

Mesophilic
aerobic
sporesa,f
Aerobic
thermophilesa,b
Yeastsa,g

Mouldsa,g

a
b
c
d
e
f
g

N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode
N
Range
Mode

Flour

Bran

Wheat germ

71
101 107
102
71
100 103
100
71
100
100
71
102 105
102
71
10 1 101
10 1
71
100 103
100
55
101 102
101
71
102 103
102
71
102 103
102

54
102 107
104
54
100 101
101
54
100 101
100
54
102 105
104
54
10 1 102
100
54
100 102
101
43
101 104
101
54
102 104
102
54
102 103
102

43
101 106
104
43
100 103
101
43
100
100
43
102 105
104
43
10 1 101
10 1
43
100 102
100
27
101 102
101
43
102 105
102
43
102 105
102

Colony forming units/gram sample.

Detection limit 10 cfu/g.
Most probable number/gram sample.
Detection limit 3 MPN/g.
Detection limit 0.3 MPN/g.
Detection limit 1 cfu/g.
Detection limit 100 cfu/g.

Yeast and moulds were frequently detected at all

stages of the milling process (Tables 2 and 3). Penicillium, Aspergillus, Aureobasidium and Cladosporium
spp. were common throughout (Table 5). Field fungi
such as Aureobasidium, Alternaria and Fusarium were
generally detected in lower numbers in end products
than on incoming wheat. Storage moulds, particularly
Eurotium, were present in higher numbers in end
product bran and wheat germ. The incidence of Wallemia sebi in conditioned wheat was higher than on
incoming wheat (Table 5), indicating that the conditioning bins and equipment may be a source of contamination.

141

Eighty-one percent of incoming wheat samples

tested positive for B. cereus (Table 1), indicating a
significant degree of field contamination. A higher
proportion of end products tested positive for B.
cereus (93% of flour and 94% of bran samples),
indicating B. cereus spores are also present in milling
equipment, contaminating otherwise clean product.
Although B. cereus was frequently detected throughout the process, a majority of samples contained less
than one spore per gram.
E. coli was detected sporadically at low levels,
usually in the early stages of milling or in end
products derived from the outer grain layers. Incoming grain rarely contained E. coli, so contamination
probably occurred during the conditioning process.
No Salmonella was detected in any end product.
Environmental serovars S. Chester and S. Hvittingvoss were detected in two milling samples (< 0.5% of
412 samples).
There was no obvious difference in microbiological quality of products from mills using older style
equipment to the newest mills in the survey (results
not shown). Similarly, geography had little impact,
except that microbiological counts in mills in Queensland, where the weather is warmer, were slightly
higher than the other mills (results not shown).
3.2. Passage of microorganisms through the milling
process
Figs. 1 3 show a comparison of the distribution of
mesophilic aerobic counts, B. cereus counts and yeast
and mould counts, observed at particular stages of the
milling process and in end products. Comparable
results (not shown) were obtained for the distribution

Table 4
The quality of flour milled from wheat with low and high mesophilic aerobic counts
Range of mesophilic
aerobic countsa
Wheat

Resulting flour

101 103
104 106

7  101 4  103
2  102 7  107

Number
of
millings

Number of flour samples

containing >104 cfu/g
mesophilic aerobic countb

28
27

0
6

cfu/g.
Proposed microbiological quality guideline for flour (refer to
Table 6).
b

142

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

Table 5
Percentage of wheat and end product samples containing identified fungal genera
Genus

Wheat before conditioning

(n = 58)

Wheat after conditioning

(n = 90)

Flour
(n = 81)

Bran
(n = 54)

Wheat germ
(n = 42)

Aureobasidium
Cladosporium
Alternaria
Fusarium
Penicillium
Aspergillus
Eurotium
Wallemia
Endomyces
Trichoderma
Epicoccum
Rhizopus
Mucor
Absidia
Paecilomyces
Scopulariopsis
Geotrichum
Trichothecium
Ulocladium
Curvularia
Drechslera

91
53
12
17
43
63
26
7
3
2
5
2
2
n.d.
n.d.
n.d
n.d.
2
2
3
2

75
39
7
2
51
60
18
27
14
4
4
4
n.d.
1
1
1
1
0
n.d.
n.d.
n.d.

15
32
9
4
65
68
22
1
4
2
2
4
2
4
n.d.
4
n.d.
n.d.
n.d.
n.d.
n.d.

52
30
15
5
57
56
33
2
7
2
n.d.a
2
n.d.
n.d.
2
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.

36
32
9
2
70
68
47
7
20
2
5
5
n.d.
n.d.
2
5
n.d.
n.d.
n.d.
2
n.d.

n.d.= not detected.

of coliforms, Bacillus spp. and mesophilic aerobic

spores, indicating that there are definite shifts in these
populations during the milling process. As outer grain
layers are removed from the grain, so too are microbial contaminants, with end product flour having the
lowest viable counts.
3.2.1. Incoming wheat
Incoming wheat provides a prime contamination
source for end products. The modal (m) aerobic
mesophilic count was 105 cfu/g, with 55% of samples
containing counts greater than or equal to 104 cfu/g, up
to a maximum of 107 (Fig. 1). Bacillus spp. (m = 104
cfu/g), yeasts and moulds (m = 103 cfu/g for both)
constitute a significant proportion of the microbial
population of incoming wheat (Table 2). Thermophiles
and coliforms, although higher in wheat than in subsequent milling samples (m = 101 cfu/g), were of minor
importance. Most Bacillus spp. entered the milling
process on incoming wheat as vegetative cells, as
mesophilic aerobic spore numbers were generally
three to four log units lower (Table 2). Mesophilic
aerobic spore counts (m = 10 cfu/g) were low on wheat
before conditioning compared with other spoilage

organisms. Mould genera entering the milling system

on wheat most frequently comprised Aspergillus, Cladosporium, Penicillium and Aureobasidium (Table 5).
E. coli and B. cereus were detected at their minimum detection level or one log unit higher in the
majority of samples. Modal E. coli (1 MPN/g) and B.
cereus (1 MPN/g) counts were low on wheat before
conditioning. Salmonella was not detected on wheat
before conditioning (Table 2).
Wheat quality is of high importance to flour
quality. A comparison of the milling of high and
low microbiological quality wheats and their respective end product flour showed wheat of low microbiological quality yields flour with high microbial
loads and vice versa (Table 4). There was also considerable variation in the quality of incoming wheat
from different geographical areas, with wheat from
hotter, wetter areas generally carrying the highest
microbial loads (data not shown).
3.2.2. Conditioning
In milling operations, water can encourage microbial growth in residues on mill machinery. Higher
mesophilic aerobic counts were more frequent after

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

143

Fig. 1. Distribution of aerobic mesophilic counts in milling process and end products showing the percentage of samples positive at each log10
value.

144

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

Fig. 2. Distribution of B. cereus in milling process and end products showing the percentage of samples positive at each log10 value.

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

145

Fig. 3. Distribution of yeasts and moulds in milling process and end products showing the percentage of samples positive at each log10 value.
Yeasts: ; moulds: n.

146

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

conditioning. Samples with 104 cfu/g or greater

increased from 56% to 73% after conditioning (Fig.
1). Although the modal count of most microbial
populations (thermophiles, coliforms, mesophilic
aerobic spores, yeast) remained similar to those in
incoming wheat, conditioning influenced the range
and maximum counts in wheat products, with the
range of E. coli counts the most affected (Table 2).
B. cereus counts generally decreased after conditioning, with the mode moving from 100 to 10 1 MPN/g
(Table 2). After conditioning, E. coli was detected in
previously uncontaminated wheat, increasing the
range detected from 100 to 102 MPN/g (Table 2).
This was probably due to transfer from poorly cleaned
conditioning bins and equipment.
Inspection of some mills during the survey
detected build-up of grain residues in conditioning
augers, elevators and storage bins. Most bacterial
species require >0.90 aw and most moulds >0.72 aw
for growth (Pitt and Hocking, 1997). As product
holding times are relatively short (8 36 h) and the
maximum aw found during conditioning was 0.75
(Fig. 4), the observed increases in microbial load were

most likely the result of contamination from encrusted

grain dust and residues inside the bins.
The populations of Bacillus spp., yeasts and
moulds were reduced after conditioning. The modal
Bacillus spp. count decreased from 104 to 102 cfu/g
after conditioning (Table 2), and the percentage of
samples containing fungal counts 103 cfu/g or greater
decreased by up to 26% (Fig. 3). The most frequent
mould genera were Penicillium, Aspergillus and Cladosporium, although the incidence of W. sebi in
conditioned wheat suggests this mould may be present
in conditioned wheat storage bins (Table 5). The low
proportion of storage moulds (6% of the total mould
population was W. sebi), however, indicates that such
mould growth is not of great concern to overall
product quality.
3.2.3. Break and reduction steps
Modal counts remained similar for many microorganisms, but the number of samples containing
high counts after break and reduction steps reduced
substantially (Table 2). The number of samples with
high B. cereus, yeast and mould counts decreased as

Fig. 4. Average water activities of milling and end products for mills located in New South Wales, Victoria, Queensland and Western Australia,
and a combined National average.

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

product passed through break, coarse and fine reduction stages (Table 2; Figs. 2 and 3). Samples with
aerobic mesophilic counts 104 cfu/g or greater and
yeast and mould counts 103 or greater decreased by
43% and 73%, respectively, after this process (Figs. 1
and 3). The maximum aerobic mesophilic count
decreased from 107 to 106, and yeast counts from
106 to 103 cfu/g after fine reduction (Table 2; Figs. 1
and 3). The maximum count observed for total
Bacillus spp. also decreased from 107 to 105 cfu/g
(Table 2). B. cereus counts were reduced after fine
reduction with counts equal to or greater than 101
decreasing by 48% (Fig. 2). E. coli was largely
unaffected by break steps, although the maximum
value detected did decrease from 102 to 101 cfu/g
(Table 2).
3.2.4. End products
Modal counts were substantially reduced for nearly
all organisms in end product flour (Table 3). The
modal aerobic mesophilic count was 102 cfu/g in flour
compared with 105 cfu/g in wheat before conditioning
(Tables 2 and 3). The mode and maximum counts for
all microorganisms decreased by one to three log units
compared with levels in early milling stages (Tables 2
and 3). The number of flour samples containing
Bacillus spp. and yeasts decreased by 26% and 55%
compared with wheat entering the mill. Mesophilic
spore former counts in flour were one log unit lower
(Tables 2 and 3). The proportion of yeast and mould
counts 103 cfu/g or greater decreased by 88% compared with wheat before conditioning.
The modal B. cereus count in end product flour
decreased by one log unit (Tables 2 and 3). One
sample of end product flour contained 9 MPN/g E.
coli (Table 3). Only small reductions were observed
between fine reduction stages and end product flour,
indicating that physical removal of microorganisms
during break stages exerts the greatest influence over
microbial levels in end product flour.
Viable aerobic mesophilic counts, coliforms, Bacillus spp. and mesophilic aerobic spores were often
higher in wheat germ than the other outer grain
fractions (Table 3). Wheat germ had lower counts of
E. coli and B. cereus than bran, however. Bran is often
heavily contaminated, possibly because it is a composite of fractions from several areas of the mill. Modal
total aerobic counts in bran and wheat germ were 104

147

cfu/g (Table 3), comparable with that of incoming

wheat (105 cfu/g) (Table 2).
3.3. Quality guidelines for Australian milling products
Microbiological guidelines for flour have been
proposed for various countries (Cicognani et al.,
1975; Spicher, 1986; Parpaiola, 1989; Potus and
Suchet, 1989; Richter et al., 1993; International Commission on Microbiological Specifications for Foods,
1998). Based on the results obtained in this survey, an
indication of the quality of Australian flour that can be
achieved with good manufacturing practice is presented in Table 6. Ninety to ninety-five percent of
samples assayed fell within these limits. Only six flour
samples contained aerobic mesophilic counts greater
than the proposed limit of 104 cfu/g. These samples
were all milled from wheat with high total aerobic
counts (Table 4).
Viable counts, coliform, yeast and mould counts
obtained in this survey are similar to those reported
from the USA and generally better than those reported
for Europe. A survey by Richter et al. (1993) found
US flour contained mean counts of 103 104 cfu/g,
depending on wheat type. Spicher (1986) reported
German flour contained mean counts of 104 cfu/g,
coliform counts of 102 cfu/g and mould counts of 103
cfu/g. Italian flour contained similar total aerobic,
coliform and fungal counts to those reported for
German flour (Cicognani et al., 1975; Ottogalli and
Galli, 1979). Potus and Suchet (1989) detected 104
cfu/g total aerobes and 103 cfu/g moulds in French
flour.
Salmonella was not detected in Australian flour
during this survey. Richter et al. (1993) reported
Salmonella in 1.32% of 3040 samples made from
Table 6
Proposed achievable microbiological quality for flour, wheat germ
and bran
Microorganism

Mesophilic aerobes
Coliforms
E. coli
B. cereus
Mesophilic aerobic spores
Yeasts and moulds

Acceptable quality limit (cfu/g)

Flour

Germ

< 10
< 102
< 10
< 10
< 102
< 103

< 10
< 103
< 10
< 10
< 103
< 104

Bran
< 105
< 103
< 10
< 10
< 102
< 104

148

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

various types of US wheat. Only one sample of

Australian flour contained E. coli, at the level of 9
MPN/g. Graves et al. (1967) found E. coli in only 1 of
16 US flour samples, but Richter et al. (1993) reported
12.8% of US flour and 22% of US durum wheat flour
contained E. coli.
The literature contains little information on the
incidence of B. cereus in flour. Our results and those
of Eyles et al. (1989) indicate B. cereus commonly
occurs in Australian wheat flour, but usually at low
levels.
3.4. Food Safety Objectives for the flour milling
industry
The International Commission on Microbiological
Specifications for Foods (ICMSF) has recently proposed a preventative approach for managing microbial
hazards in foods (International Commission on Microbiological Specifications for Foods, 2000). The
approach uses the concept of a Food Safety Objective
(FSO) as a functional link between risk assessment
and risk management. An FSO is the maximum
concentration and/or frequency of hazard in a food
that, at the time of consumption, will provide an
appropriate level of public health protection.
One advantage of this approach is that it delivers the
desired outcome of control measures that may be
applied throughout the food chain, but still allows
flexibility in their application. Control measures
include controlling the critical level of a hazard, reducing the level of a hazard and preventing an increase of a
hazard. The approach provides a framework for comparing equivalency of different control measures. For
example, if the critical level of a hazard in a food is low,
then subsequent control measures may be milder.
The data generated in this work can form the basis
for setting critical levels and for the development of
control measures to be applied in HACCP systems for
flour and other mill products to meet a given FSO.

4. Conclusions
The microbiological quality of the incoming wheat
has a strong bearing on the ultimate quality of milling
end products. Flours with higher microbial counts
were usually derived from wheat of poorer micro-

biological quality. As wheat grain layers are separated,

surface-adhering contaminants are concentrated in end
product bran, wheat germ and pollard. The inner
semolina fraction contains low microbial counts, and
end product flour is the cleanest end product of the
milling process.
Variables within the milling process and residue
build up in milling plant and equipment also constitute
a significant source of microbiological contamination.
Increases in microbial levels in certain midstream
products indicate that in some mills, microorganisms,
particularly spore-forming species, may reside in
equipment, adversely affecting end product quality.
Attention to mill hygiene is an important factor in the
production of flour and other fractions with low microbial counts. The quality of Australian flour compares
favourably with that of the USA and European nations.

Acknowledgements
This project was funded by the Quality Wheat
Cooperative Research Centre. The participation in the
milling survey and generous assistance of GoodmanFielder Milling and Baking, Bunge-Defiance Milling
and Weston Milling is greatly appreciated. The
authors thank Ms. Cathy Moir and Ms. Nancy Jensen
for critical review of this manuscript.

References
Anonymous, 1958. A search for organisms of the Salmonella group
in wheat and flour. Montana Bulletin of the Ministry of Health
and Public Health Laboratory Service 41, 88 91.
Cicognani, G., Pedretti, C., Cerrato, A., 1975. Caratteristiche microbiologiche delle farine di frumento (Microbiological characteristics of wheat flour). Industrie Alimentari 14 (7/8), 60 64.
Dack, G.M., 1961. Public health significance of flour bacteriology.
Cereal Science Today 6, 9 10.
Eyles, M.J., Moss, R., Hocking, A.D., 1989. The microbiological
status of Australian flour and the effects of milling procedures
on the microflora of wheat and flour. Food Australia 41,
704 708.
Graves, R.R., Rogers, R.F., Lyons, A.J., Hesseltine, C.W., 1967.
Bacterial and actinomycete flora of Kansas Nebraska and Pacific North-West wheat and wheat flour. Cereal Chemistry 44,
288 299.
Hesseltine, C.W., Graves, R.R., 1966. Microbiology of flours. Economic Botany 20, 156 168.

L.K. Berghofer et al. / International Journal of Food Microbiology 85 (2003) 137149

International Commission on Microbiological Specifications for
Foods, 1998. Microorganisms in Foods: 6 Microbial Ecology
of Commodities. Blackie Academic and Professional, London.
Chapter 8.
International Commission on Microbiological Specifications for
Foods, 2002. Microorganisms in Foods: 7 Microbiological Testing in Food Safety Management. Kluwer Academic/Plenum
Publishers, New York.
Ottogalli, G., Galli, A., 1979. Microbiological quality of flours:
sour dough for bakery products and spaghetti. In: Jarvis, B.,
Christian, J.H.B., Michener, H. (Eds.), Food Microbiology and
Technology. Proceedings of the International Meeting on Food
Microbiology and Technology, Tabiabo (Parma) Italy, April
20 23, 1978, Medicinia Viva, Parma, Italy, pp. 141 153.
Parpaiola, D., 1989. Affidabilita` del controllo di qualita` in unindustria molitoria (Reliability of quality control in a flour milling
industry). Tecnica Molitoria 40, 681 691.
Pitt, J.I., Hocking, A.D., 1997. Fungi and Food Spoilage, 2nd ed.
Aspen Publishing, Gaithersburg, MD, USA.
Potus, J., Suchet, P., 1989. Les proble`mes de microbiologie en
meunerie (Microbiological problems in milling). Industries des
Cereales 58, 27 33.
Richter, K.S., Dorneanu, E., Eskridge, K.M., Rao, C.S., 1993. Microbiological quality of flours. Cereal Foods World 38, 367 369.
Spicher, G., 1986. Merkpunkte fur die Beurteilung der mikrobiologisch-hygienischen Qualitat von Weisenmehlen (Judging the
microbiological-hygienic quality of white flour). Die Muhle &
Mischfuttertechnik 33, 449.

149

Standards Australia, 1991a. Australian Standard, Food Microbiology. 1766.2, Examination for specific organisms. 1766.2.1:
Standard plate count.
Standards Australia, 1991b. Australian Standard, Food Microbiology. 1766.1, General procedures and techniques. 1766.1.3: Colony count pour plate method.
Standards Australia, 1991c. Australian Standard, Food Microbiology. 1766.1, General procedures and techniques. 1766.1.4: Colony count surface spread method.
Standards Australia, 1991d. Australian Standard, Food Microbiology. 1766.2, Examination for specific organisms. 1766.2.6:
Bacillus cereus.
Standards Australia, 1991e. Australian Standard, Food Microbiology. 1766.2, Examination for specific organisms. 1766.2.5: Salmonellae.
Standards Australia, 1992. Australian Standard, Food Microbiology. 1766.2, Examination for specific organisms. 1766.2.3:
Coliforms and Escherichia coli.
Standards Australia, 1997. Australian Standard, Food Microbiology. 1766.2, Examination for specific organisms. 1766.2.2
(1997) Colony count of yeasts and moulds.
Stevenson, K.E., Segner, W.P., 1992. Mesophilic aerobic sporeformers. In: Vanderzant, C., Splittstoesser, D. (Eds.), Compendium of Methods for the Microbiological Examination of Foods,
3rd ed. American Public Health Association (APHA) Technical
Committee on Microbiological Methods for Foods. American
Public Health Association, Washington, DC, pp. 265 274.
Chapter 18.