Journal of Cereal Science 25 (1997) 103–110

Genetic and Agronomic Variation in Arabinoxylan

and Ferulic Acid Contents of Durum Wheat
(Triticum durum L.) Grain and Its Milling Fractions
I. Lempereur, X. Rouau and J. Abecassis

INRA-Unité de Technologie des Céréales 2 Place Viala, 34060 Montpellier cedex 01, France

Received 21 December 1995

ABSTRACT
Total arabinoxylan (AXt), water-extractable arabinoxylan (WeAX) and ferulic acid (FA) from five
cultivars of durum wheat (Triticum durum L.), grown under four different agronomic conditions, were
measured. Among the varieties analysed, AXt, WeAX and FA contents ranged between 4·07%–6·02%,
0·37%–0·56% and 0·784 mg/g–7·98 mg/g, respectively. High genetic and agronomic variability was
detected for AXt, WeAX and FA. AXt and FA increased sharply in milling products for extraction
rates above 60%. FA was quantified in durum wheat milling fractions. High concentrations of FA
esterified to cell-wall arabinoxylans were found in the aleurone layer (69% of total FA), germ and
seedcoat (26·6% of total FA). Only trace amounts were detected in the starchy endosperm (1·4% of
total FA). A highly significant correlation appeared between AXt and FA contents.
 1997 Academic Press Limited

Keywords: durum wheat, ferulic acid, arabinoxylans, milling.

INTRODUCTION extractable arabinoxylan (WeAX) contents in

whole common wheat (Triticum aestivum L.) grain
Arabinoxylans are located in the cell walls of are 6·7% and 0·70%, respectively5. Similar values
cereals. Their main structural feature is the pres- were found for durum wheat (Triticum durum L.)6.
ence of a xylan backbone consisting of b-1,4 D- Studies focusing on the arabinoxylans of durum
xylopyranose units some of which are substituted wheat revealed that durum wheat arabinoxylans
by a-L-arabinose units at O-31,2 oxygen or at have higher levels of arabinose, indicating a more
O-2 and O-33,4 positions. Ferulic acid (FA) is highly substituted structure. They also have a
esterified at O-5 position of arabinose residues. higher molecular weight. WeAX from durum
Some arabinoxylans are water-extractable. wheat endosperm are just as branched as those
Average total arabinoxylan (AXt) and water- from common wheat6. Other cereals like barley
exhibit AXt contents in whole grain similar to
those in common wheat7. But in the starchy en-
 : FA=ferulic acid; AXt=total ara- dosperm, the proportion of AXt is much lower
binoxylans; WeAX=water-extractable arabinoxylans; (20%8 vs 72%9). The arabinose:xylose ratio is
WuAX=water-unextractable arabinoxylans; WeAX/ higher for barley (0·7210) than for wheat (0·5011).
WuAX=water-extractable/water-unextractable ara-
binoxylan ratio; TMCA=2,5-trimethoxy-trans-cin-
Rye contains higher amounts of AXt and WeAX
namic acid; HPLC=high performance liquid than other cereals12,13.
chromatography; d.m.=dry matter; TGW=1000 The study of the histological distribution of AXt
grain weights; W1=agronomic treatment inducing showed that they increase from the centre to
water stress; W2=agronomic treatment with irrigation; the periphery of endosperm. They are mainly
N1=agronomic treatment inducing nitrogen stress; concentrated in the seedcoats of wheat, barley and
N2=agronomic treatment with nitrogen fertilisation. rye. Milling by-products contain 9·9–28·4% AXt

0733–5210/97/020103+08 $25.00/0/jc960090  1997 Academic Press Limited

104 I. Lempereur et al.

and 1·29–1·68% WeAX in Triticum aestivum. In Milling fractions

common wheat flour, mean levels of AXt range
from 1·50% to 2·10% and water extractable from Milling was carried out on a semi-industrial scale
0·70 to 0·80%14. AXt contents of durum wheat semolina production mill (150 kg/h) at the Cereal
(Triticum durum L.) flour or semolina ranged from Technology Unit (INRA-Montpellier, France). It
1·80% to 2·80%15. Also, the WeAX content de- included a cleaning unit (cleaner, separator, sorter
creases from the inner to the outer part of the and beater), a grain conditioning and storage unit,
grain. The extractability of arabinoxylan shows and a milling unit (four break rolls, four sizing
similar trends, and is especially low in the seed- rolls, three plansifter stacks and three double-deck
coats. purifiers).
Ferulic (4-hydroxy-3-methoxycinnamic) acid Clean wheat grain was tempered to 15·0% H2O
(FA) is the most common phenolic acid in mono- for 15 h, and then at 17·0% for 3 h just prior to
cotyledonous cell walls16. In wheat, this acid is milling. After milling, the following 18 fractions
esterified to arabinose units of cell-wall ara- were obtained: six purified semolinas, four break
binoxylans, and occurs in high concentrations in flours, four sizing flours, and four middlings: coarse
aleurone, pericarp and embryo cell walls, but only brans, purified fine brans, sized fine bran and
in trace amounts in the starchy endosperm of ripe shorts. Milling yields were expressed in terms of
kernels17. The trans isomer is predominant, and dry matter. Samples were collected during milling
accounts for 90% of the total phenolic acids in and stored at −4 °C.
common flour18.
To date, very little information is available on Chemical analysis
the genetic contribution and the effect of the
cultivation conditions on the arabinoxylan and Moisture content was determined by ISO 711-
phenolic acid contents of durum grain. The present 1978 method by heating at 130 °C for 2 h. Ash
study was aimed at investigating genetic and agro- content was determined by the ISO 2171-1980
nomic variability with respect to AXt and WeAX standard procedure at 900 °C for fractions with
and FA contents in durum wheat. Changes in low mineral levels and at 550 °C for middlings. For
these levels in milling fractions were also monitored the analyses described below, the milling fractions
in order to analyse the separation of various histo- were ground in a Cyclotec 1093 Sample Mill
logical fractions during milling. (Tecator, Höganas, Sweden) with a 1·0 mm mesh
sieve. The analytical results were expressed as dry
matter.

Total and water-extractable arabinoxylan

MATERIALS AND METHODS analysis

Wheat samples For AXt, ground milling fractions (0·1 g) were

weighed in a 10 mL Pyrex test tube. After adding
The durum wheat varieties Agridur, Ambral, Ar- 1·0 M sulphuric acid (5·0 mL), the sample was
dente, Cando and Primadur cultivars were grown suspended and stirred with a magnetic bar. The
at two separate crop sites (Melgueil and Auzeville, suspension was heated to 100 °C for 10 min for
southern France) at INRA (Institut National de flours, semolinas and whole grains or 30 min for
la Recherche Agronomique) under four different middlings. The test tubes were cooled on ice and
agronomic conditions. Water stress was in- then centrifuged for 15 min at 5000 g. An aliquot
vestigated at Melgueil: batch W1 was rep- of the clear supernatant fraction was collected and
resentative of an unirrigated Mediterranean crop; diluted for colorimetric analysis with an auto-
the irrigated batch (W2) was sprinkled with a total analyser.
supply of 102 mm water from the onset to the end For WeAX, ground milling fractions (1·0 g for
of the heading phase. Nitrogen was varied at flours, semolinas and whole grains, and 0·5 g
Auzeville: batch N1 only received 52 kg/ha of for middlings) were weighed into a 10 mL tube.
nitrogen over the cropping period; the control Ultrapure water (4·0 mL flour, semolina, whole
batch (N2) was fertilised with 102 kg/ha of nitrogen grain; 5·0 mL middlings) was added. Mixtures
over the same period. were homogenised and shaken in a rotary shaker
 Arabinoxylan and ferulic acid in durum wheat 105

(40 rev/min) for 15 min at 20 °C. After cent- conditions tested (Table I). AXt and WeAX con-
rifugation (5 min, 5000 g), the supernatants were tents ranged from 4·07% to 6·02% and from
collected and diluted for analysis with an auto- 0·37% to 0·56%, respectively. For whole common
analyser. AXt and WeAX were analysed using a wheat grain the mean AXt and WeAX are 6·70%
semi-automatic method19 adapted from the Doug- and 0·70%5, respectively.
las20 phloroglucinol technique. Water un- Extractability was characterised by WeAX/
extractable arabinoxylam (WuAX) was calculated WuAX ratio. It was found to be highest for cvs.
by difference. Ambral, Cando and Agridur, while cvs. Ardente
and Primadur had lower but similar ratios (10·1%
Analysis of esterified ferulic acid for Primadur and 13·1% for Ardente), even though
their AXt contents differed twofold. This indicated
Ground samples (0·1 g) were defatted with hexane that the extractability of arabinoxylans was in-
(5·0 mL, 35 °C) for 1 h with rotary shaking (40 dependent of the AXt content. Varieties with
rev/min). After centrifugation (5 min, 3800 g, high AXt contents also had high levels of WeAX.
20 °C) the defatted pellet was dried for 24 h at Similar values are generally found for common
room temperature. wheat (Triticum aestivum)5.
Ground defatted semolina samples (0·1 g) were AXt and WeAX contents were found to be
saponified for 2 h in the dark with 2·0 M NaOH lower under the nitrogen conditions tested at Auze-
(10·0 mL) at 35 °C (with shaking, in the presence ville (N1 and N2) than the results obtained under
of argon). Internal standard, (2,5 trimethoxy-trans- the water stress conditions tested at Melgueil (W1
cinnamic acid (TMCA)21 was then added, and the and W2). AXt contents differed significantly be-
solution was adjusted to pH 2·0 with 4·0 M HCl. tween the two crop sites. The levels were higher
The phenolic acids were extracted twice with ether in wheat grown in the Mediterranean climatic
(5 mL) with shaking and 5 min centrifugation at zone23. The extractability patterns were in line
3800 g. The ether phases were collected in amber with this observation, with results from the W1
test tubes and evaporated to dryness in the pres- treatment close to those obtained with the N1 and
ence of argon. N2 treatments. Nitrogen stress (treatment N1) did
The dry extract was dissolved in methanol not modify the AXt and WeAX contents. AXt
(2·0 mL), injected on an RSil C18 column (250 mm were higher in case of the water stress conditions
× 4·6 mm) (Interchim, Paris, France) and eluted (treatment W1), while WeAX contents were only
at 1·0 mL/min with the following gradient profile: slightly modified. A comparison of treatments W2
10–40% B for 30 min; 40% B for 5 min and then and N2 revealed a significant site-specific effect
40–10% B for 10 min with solvent B (acetonitrile) on AXt, whereas it was insignificant on WeAX.
and solvent A (0·05 M acetate buffer, pH 4, 35 °C). AXt and WeAX contents and WeAX/WuAX ra-
Detection was at 320 nm22. tios were mainly affected by the wheat genotype
as shown by the genetic effect/treatment ratios
Statistical analysis (Table I).
There was a very high genetic variability with
The Statgraphics software package (Manugistics,
respect to the FA contents of the durum wheat
Rockville, U.S.A.) was used for the statistical ana-
varieties studied. The concentrations ranged from
lyses. The effects were studied by ANOVA. The
0·693 mg/g d.m. to 2·443 mg/g d.m. and the mean
coefficients of variation for the arabinoxylan and
levels were higher than those reported for common
FA measurement procedures were 3% and 4%,
wheat, which ranged from 0·5 mg/g d.m.22 to
respectively. Samples were analysed in duplicate
0·982 mg/g d.m.24.
and results expressed as mean values.
The Auzeville treatments led to lower FA con-
tents than those obtained with the Melgueil treat-
RESULTS AND DISCUSSION ments. The nitrogen treatment did not alter the
Variation in arabinoxylan and ferulic acid FA contents, whereas the water stress treatment
contents of whole durum wheat grain: varietal (W1) increased the levels significantly. The genetic
effect was more important than the agronomic
and agronomic effects
effect.
FA, AXt and WeAX contents varied between the The FA/AXt ratio reflected the extent to which
durum wheat cultivars studied and the agronomic FA were esterified to the cell-wall arabinoxylans.
106 I. Lempereur et al.

Table I Mean values of total (AXt) and water-extractable (WeAX) arabinoxylan, water-extractable/water-unextractable
arabinoxylan ratio (WeAX/WuAX), ferulic acid (FA) and ash contents for different wheat cultivars and agronomic conditions
(N1, N2, W1 and W2). Analysis of the genotypic (G) and agronomic (A) contribution of the residue and of the ratio genotypic/
agronomic (s2G/s2A)
AXt WeAX WeAX/ FA mg/g FA per ash
% d.m. % d.m. WuAX % d.m. AXta %d.m.
Ambral 5·14 0·44 9·4 1·98 38·3 2·03
Primadur 6·02 0·56 10·1 1·30 21·5 2·06
Cando 4·83 0·41 9·3 1·43 29·4 2·03
Agridur 4·76 0·37 8·4 1·06 22·3 1·86
Ardente 4·07 0·46 13·1 0·78 19·4 2·09
N1 4·77 0·44 10·5 1·12 23·5 1·91
N2 4·60 0·44 10·9 1·12 24·3 1·96
W1 5·36 0·49 10·2 1·60 29·6 2·08
W2 5·14 0·42 8·8 1·40 27·4 2·10
Gb 73·0∗∗ 58·8∗∗ 54·0∗∗ 73·3∗∗ 78·1∗∗ 33·6 ns
Ab 16·0∗ 11·9 ns 13·5 ns 18·8∗∗ 9·9 ns 33·4 ns
Residueb 10·9 29·3 32·5 7·9 12·0 32·9
s2G/s2Ac 4·5 4·9 4·0 3·9 6·5 1·0
a
lg FA/mg AXt.
b
Percentage of genotypic (G), agronomic (A) variation and of residue.
c
Ratio of genotypic (G) influence verse agronomic (A).
∗P<0·05.
∗∗P<0·01.
ns=non significant.

The high genetic variability in the FA content central and intermediate endosperm were very
was responsible for changes in this ratio, but no pure, and were produced at the head of the milling
agronomic effect was noted. There were no no- process and first four purifiers. Their mean FA levels
ticeable agronomic or genetic effects on the ash were 140 lg/g d.m. for cv. Primadur and 120 lg/
contents. g d.m. for cv. Ardente. The same patterns were
The 1000-grain weights for the durum wheat noted for AXt contents. Above 40% extraction, FA
varieties studied were found to range from 25·3 g contents increased slowly. The products were de-
to 42·9 g, thus affecting the kernel/seedcoat ratio rived from the intermediate endosperm and slightly
and the arabinoxylan and FA contents. Small grain contaminated with aleurone fragments.
varieties like Primadur (25·3 g TGW) exhibited Above 60–70% extraction, the highest bran/
arabinoxylan and FA contents higher than large shorts carryover in the products occurred: a slight
grain varieties like Ardente (42·9 g TGW) (Table variation in the extraction rate caused substantial
I). Therefore, ‘per grain’ contents did not differ variations in FA and AXt levels. These variations
from the above results. could be explained by a concentration gradient in
the endosperm (for AXt), seedcoat friability and
breakdown of the aleurone layer. The proportion
Ferulic acid and total arabinoxylan contents in of endosperm in the grain and the potential peak
relation to extraction rates extraction rate were determined along the abscissa
Cumulative content curves were plotted for FA at the breakpoint of the slope. This value provided
and AXt contents as a function of extraction an estimate of semolina milling values for the
rates for cvs. Ardente and Primadur for the W2 different varieties relative to the extraction rate.
agronomic treatment (Figs 1 and 2). No data Cultivars Primadur and Ardente had potential
concerning WeAX are reported in this section, milling yields of 80% and 82%, respectively.
considering that AXt and FA are adequate markers Above 80% extraction, milling products were
for the wheat grain cell-wall material. derived from the bran/shorts layers. Con-
The FA content varied slightly within the 0–40% sequently, there were almost no variations in FA
extraction rate range. Semolinas derived from the and AXt levels, and the curve was linear.
 Arabinoxylan and ferulic acid in durum wheat 107

0.9 6

0.8
5
0.7

Total arabinoxylan (% d.m.)

Ferulic acid (mg/g d.m.)

0.6 4

0.5
3
0.4

0.3 2

0.2
1
0.1

0
0 20 40 60 80 100
Extraction rate (% d.m.)

Figure 1 Changes in (Φ) ferulic acid (lg/g) and (Ε) total arabinoxylan (%) contents as a function of extraction rate for cv.
Ardente (agronomic condition W2).

1.4 8

1.2 7

6
Total arabinoxylan (% d.m.)

1.0
Ferulic acid (mg/g d.m.)

5
0.8
4
0.6
3

0.4
2

0.2 1

0
0 20 40 60 80 100
Extraction rate (% d.m.)

Figure 2 Changes in (Φ) ferulic acid (lg/g) and (Ε) total arabinoxylan (%) contents as a function of extraction rate for cv.
Primadur (agronomic condition W2).
108 I. Lempereur et al.

Table II Ferulic acid (FA), total arabinoxylan (AXt) and ash contents of different mill streams from Ardente and Primadur
cultivar expressed as a proportion (%) of the values for whole grain (agronomic conditions W2)

Ardente Primadur

FA Axt ash FA Axt ash

Whole grain 100·0 100·0 100·0 100·0 100·0 100·0
Inner semolina 18·4 36·3 37·8 11·0 38·0 39·8
Inner flour 47·9 53·0 104·1 39·0 59·0 100·6
Intermediate semolina 22·5 31·8 46·8 18·7 53·1 48·8
Intermediate flour 55·6 50·9 129·5 44·3 74·3 115·0
Peripheral semolina 76·2 95·7 89·4 34·8 70·7 68·7
Peripheral flour 75·8 72·0 152·7 53·4 74·4 142·4
Shorts 376·5 570·9 281·1 263·3 340·6 218·5
Coarse bran 413·7 678·8 280·2 432·7 475·8 256·5

Distribution of ferulic acid and total for the endosperm, 1·4% of the FA was present
arabinoxylans in the different milling fractions in 80–84% of the total grain.
Variations in FA, AXt and ash contents in the
Table II gives the FA, AXt and ash contents of endosperm can be compared in Table II. These
the different milling fractions with cvs. Ardente and levels increased from the centre to the outer layer of
Primadur (W2 agronomic treatment), expressed as the endosperm, but there were marked differences
percentages of contents of the whole grain. with respect to the different compounds and wheat
The bran/shorts fractions were found to contain varieties studied. Increases in AXt were not
higher FA, AXt and ash contents than those from
matched by concomitant increases in WeAX. Ara-
the central endosperm (flours and semolinas). The
binoxylan extractibility therefore decreased from
milling process concentrated FA, arabinoxylans
the centre to the outer layer of the endosperm, as
and ash in the bran/shorts fractions, in agreement
also reported in common wheat5.
with previous studies on other cereals22,24.
In the same histological fractions, semolina had
In addition, the ratio of endosperm AXt (central
semolina) to grain AXt was found to be equal to lower FA, AXt and ash concentrations than flour.
the endosperm/grain ash ratio, i.e. 0·36 for cv. In cvs. Ardente and Primadur, FA and ash levels
Ardente and 0·38 for cv. Primadur. The central were found to differ by 2·5-fold in central en-
endosperm/grain FA ratio was lower, i.e. 0·18 for dosperm semolinas and flours, while AXt levels
cv. Ardente and 0·11 for cv. Primadur. differed by 1·5-fold. However, ash levels did not
The concentration gradients in semolina derived differ for flours and semolinas derived from central
from the central endosperm and the outer layer endosperm fractions.
of the grain (coarse bran) varied according to the This observation could be explained by bran
compound studied and the wheat variety. The friability, which can result in contamination of
concentration gradient between the central en- central endosperm semolinas and flours during
dosperm and the seedcoat was higher for AXt the milling process. Density classification through
(642 for cv. Ardente and 438 for cv. Primadur) purifiers eliminated contaminating bran fragments
than for FA (400) and ash (230). The FA contents from semolinas, whereas grading of flours ac-
of the different durum wheat grain tissues were cording to particle size through a sifting operation
determined after manual dissection. The results was ineffective. For fractions from the peripheral
are presented in Table III. endosperm layers, there was a higher proportion
In agreement with Fulcher’s findings for com- of aleurone cells in both semolina and flour. It
mon wheat17, FA was found concentrated in the was thus difficult to determine whether there was
aleurone layer (69%) and pericarp (29%). Only increased contamination of very small seedcoat
2% of the total FA was detected in the endosperm fragments (which are hardly extracted by puri-
and germ (Table III). This concentration effect is fication), or whether such fragments are common
emphasised by the fact that 69% of the FA was to such fractions.
distributed in 6–7% of the total grain. Conversely, A matrix of correlations between the different
 Arabinoxylan and ferulic acid in durum wheat 109

Table III Ferulic acid (mg/g dm) contents of different histological tissues (endosperm, aleurone layer, pericarp and germ)
and of whole grain of cv. Ardente
Grain Endosperm Aleurone Pericarp Germ
a
FA (mg/g d.m.) 0·87 0·17 8·82 3·71 0·08
Percent of whole grainb 100 80–84 6–7 7·5–9·5 2·5–3·5
Concentration in FA (% of 100·0 1·4 69·0 29·0 0·6
whole grain)
a
mean of duplicates.
b
From ref. 29. Y. Pomeranz (ed.) Chemical composition of kernel structures. In ‘Wheat: Chemistry and Technology’, American
Association of Cereal Chemists, St Paul, MN. (1988) pp 119.

Table IV Correlation coefficients (r) between ferulic acid There was a poor correlation between ash and
(FA), total arabinoxylan (AXt) and ash. FA contents in the central endosperm. In common
FA AXt Ash wheat, ash contents in this fraction accounted for
20–26%25 of total ash, while the FA content was
All mill streams (n=304) less than 5%. Similar results were also reported
FA 1 for durum wheat26. The correlation between the
AXt 0·97∗∗ 1
Ash 0·89∗∗ 0·90∗∗ 1 FA and ash contents in this case was higher for
Semolina and flour reduction flours (r=0·89) than for break flours
streams (n=204) (r=0·60).
FA 1
AXt 0·78∗∗ 1
Ash 0·26∗∗ 0·46∗∗ 1 CONCLUSION
Coarse bran, fine bran
and shorts streams FA and AXt levels in whole grain of durum wheat
(n=64) (Triticum durum) were especially affected by the
FA 1
AXt 0·69∗∗ 1 genotype and crop site, but very slightly by agro-
Ash 0·84∗∗ 0·52∗∗ 1 nomic conditions. These components appeared to
be good markers for bran/shorts fractions, and
∗∗P<0·01 therefore of the endosperm/seedcoat ratio. The
external histological layers (aleurone and pericarp)
showed high FA and arabinoxylan concentrations.
components studied is given in Table IV. Milling The higher levels of FA and AXt in the bran/
fraction subgroups were distinguished: endosperm shorts fractions than the central endosperm frac-
fractions (flours and semolinas) and bran/shorts tions were more marked than for ash levels in the
fractions. same fractions. The cumulative content curves
For all milling fractions, there were significant plotted for FA as a function of the extraction
correlations (at P=0·01) between AXt and FA rate highlighted the absence of a concentration
contents, between FA and ash contents and be- gradient for this component in the endosperm.
tween ash and AXt contents, with correlation There was a more substantial increase in AXt,
coefficients of 0·97, 0·89 and 0·90, respectively. indicating the presence of an arabinoxylan con-
For the starchy endosperm fractions (flours and centration gradient in the endosperm. In this re-
semolinas), there were significant correlations (at spect, the slope breakpoints were very easily
P=0·01) between FA and AXt contents (r=0·78). determined, distinguishing the starchy endosperm,
However, the correlations between FA and ash aleurone layer and seedcoat. The potential milling
contents and between AXt and ash contents (r= yield can be partially assessed, i.e. the endosperm/
0·26 and 0·46, respectively) were very low. seedcoat ratio, from this point.
For the bran/shorts fractions (middlings), there The present results with durum wheat could be
was a significant correlation (at P=0·01) between used, as with common wheat27,28, in designing a
ash and AXt contents (r=0·69) and between ash rapid technique to evaluate the purity of wheat
and FA contents (r=0·84). The FA content was millstreams. The different histological fractions
highly correlated with AXt in all fractions. (endosperm, pericarp and aleurone) could thus
110 I. Lempereur et al.

be quantified by fluorescence microscopy analysis tionated (milled) rye products. Cereal Chemistry 66 (1989)
standardised with FA concentrations. 107–111.
13. Bengtsson, S. and Aman, P. Isolation and chemical
characterization of water soluble arabinoxylans in rye
grain. Carbohydrate Polymers 12 (1990) 267–277.
Acknowledgements 14. Ciacco, C.F. and D’Appolonia, B.L. Characterization of
This study was carried out in the framework of a gelling capacity of water extractable pentosans and their
gelling capacity. Cereal Chemistry 59 (1982) 163–166.
program of the French Institut de Recherches Tech- 15. D’Appolonia, B.L. and MacArthur, L.A. Composition of
nologiques Agro-alimentaires des Céréales (IRTAC) starch, pentosans and sugar of some conventional height
focusing on milling and semolina potentials of wheats. and semidwarf Hard Red Spring wheat flours. Cereal
We are very grateful to Mrs B. Mahaut (Institut Tech- Chemistry 53 (1975) 871–873.
nique des Céréales et Fourrages, ITCF) for performing 16. Smith, M.M. and Hartley, R.D. Occurrence and nature
the physical grain analyses. of ferulic acid substitution of cell wall polysaccharides
in graminaceous plants. Carbohydrate Research 118 (1983)
65–80.
REFERENCES 17. Fulcher, R.G. Fluorescence microscopy of cereals. In
‘Food Microstructure’, Canada, Ottawa (1982) pp 167–
1. Perlin, A.S. Structure of the extractable pentosans of 175.
wheat flour. Cereal Chemistry 28 (1951) 370–381. 18. Sosulski, F., Krygier, K., and Hogge, L. Free esterified
2. Montgomery, R. and Smith, F. The carbohydrates of the and extractable-bond phenolic acids. III. Composition
Graminaceae. VII. The constitution of water extractable of phenolic acids in cereal and potato flours. Journal of
hemicellulose of endosperm of wheat (Triticum vulgare). Agricultural and Food Chemistry 30 (1982) 337–340.
Journal of American Chemical Society 77 (1955) 3325–3328. 19. Rouau, X. and Surget, A. A rapid semi-automated
3. Hoffman, R.A. Structural characterisation from white method for the determination of total and water-ex-
wheat flour. Ph.D. Utrecht (1991). tractable pentosans in wheat flours. Carbohydrate Polymers
4. Gruppen, H., Hamer, R.J. and Voragen, A.G.J. Water- 24 (1994) 123–132.
unextractable cell wall material from wheat flour. I. 20. Douglas, S.G. A rapid method for the determination of
Fractionation of alkali-extracted polymers and com- pentosans in wheat flour. Food Chemistry 7 (1981) 139–145.
parison with water-extractable arabinoxylans. Journal of 21. Scalbert, A., Monties, B., Lallemand, J.Y., Guittet, E.
Cereal Science 16 (1992) 51–67. and Rolando, C. Ether linkage between phenolic acids
5. Hashimoto, S., Shogren, M.D. and Pomeranz, Y. Cereal and lignin fractions from wheat straw. Phytochemistry 24
pentosans: Their estimation and significance. I. Pentosans (6) (1985) 1359–1362.
in wheat milled products. Cereal Chemistry 64 (1987) 30–34. 22. Pussayanawin, V. and Wetzel, D.L. High-performance
6. Medcalf, D.G., D’Appolonia, B.L., and Gilles, K.A. Com- liquid chromatographic determination of ferulic acid in
parison of chemical composition and properties between wheat milling fractions as a measure of bran con-
Hard Red Spring and Durum wheat endosperm pen- tamination. Journal of Chromatography 391 (1987) 243–255.
tosans. Cereal Chemistry 45 (1968) 539–549. 23. Hong, B.H., Rubenthaler, G.L. and Allan, R.E. Wheat
7. Henry, R.J. Pentosan and (1–3), (1–4)-b-glucan con- pentosans. I Cultivar variation and relationship to kernel
centrations in endosperm and wholegrain of wheat, bar- hardness. Cereal Chemistry 66 (1989) 369–373.
ley, oats and rye. Journal of Cereal Science 6 (1987) 253–258. 24. Rybka, K., Sitarski, J. and Raczinska-Bojanowska, K.
8. Ballance, G.M. and Manners, D.J. Structural analysis Ferulic acid in rye and wheat grain and grain dietary
and enzymic solubilization of barley endosperm cell walls. fiber. Cereal Chemistry 70 (1993) 55–59.
Carbohydrate Research. 61 (1978) 107–118. 25. Hinton, J.J.C. The distribution of ash in the wheat kernel.
9. Fincher, G.B. and Stone, B.A. Cell walls and their com- Cereal Chemistry 36 (1959) 19–31.
ponents in cereal grain technology. In ‘Advances in 26. Symons, S.J. and Dexter, J.E. Relationship of flour aleur-
Cereal Science and Technology’ Vol 8 (Y. Pomeranz, one fluorescence to flour refinement for some Canadian
ed.) American Association of Cereal Chemists, St Paul Hard common wheat classes. Cereal Chemistry 70 (1993)
MN, U.S.A. (1986). 90–95.
10. Gruppen, H. Marseille, J.P., Voragen, A.G.J., Hamer, 27. Jensen, S.V.A., Munck, L. and Martens, H. The botanical
R.J. and Pilnik, W. Mild isolation of water-insoluble cell constituents of wheat and wheat milling fractions. I-
wall material from wheat flour: composition of fractions Quantification by autofluorescence. Cereal Chemistry 59
obtained with emphasis on non-starch polysaccharides. (1982) 477–484.
Journal of Cereal Science 9 (1989) 247–260 28. Symons, S.J. and Dexter J.E. Computer analysis of fluor-
11. Vietör, R.J., Angelino, S.A.G.F. and Voragen, A.G.J. escence for the measurement of flour refinement as de-
Structure features of arabinoxylans from barley and malt termined by flour ash content, flour grade color and
celle wall material. (1992) Journal of Cereal Science 15 (1992) tristimulus color measurements. Cereal Chemistry 68 (1991)
223–236. 454–460.
12. Delcour, J.A., Vanhamel, S. and De Geest, C. Physico- 29. Pomeranz, Y. (ed.) Chemical composition of kernel struc-
chemical functional properties of rye nonstarch poly- tures. In ‘Wheat: Chemistry and Technology’, American
saccharides. I. Colorimetric analysis of pentosans and Association of Cereal Chemists, St Paul, MN, U.S.A.
their relative monosaccharide compositions in frac- (1988) pp 119.