J. Agric. Food Chem.

2005, 53, 48434852

4843

Study of the Main Constituents of Some Authentic Hazelnut

Oils
COLIN CREWS,*, PATRICK HOUGH, JOHN GODWARD, PAUL BRERETON,
MICHELLE LEES, SEBASTIEN GUIET, AND WILFRIED WINKELMANN#
Central Science Laboratory, Sand Hutton, York YO41 1LZ, United Kingdom; Eurofins Scientific,
Rue Pierre Adolphe Bobierre, B.P. 42301, 44323 Nantes Cedex 3, France, and Eurofins/
Wiertz-Eggert-Jorissen, Stenzelring 14 b, D-21107 Hamburg, Germany

This paper describes the composition of authentic hazelnut oils obtained from nuts collected from
five countries that are major suppliers of hazelnut oil. Oils were analyzed using standard methods
for fatty acids, fatty acids in the triacylglycerol 2-position, tocopherols and tocotrienols, triacylglycerols,
sterols, steradienes, and iodine value. The results were generally in good agreement with those of
other publications. Tocotrienols, previously unreported in hazelnut oil, were detected in one sample.
There were no major differences in the composition of oils from different countries. Roasting the
nuts prior to pressing had little effect on oil composition.
KEYWORDS: Hazelnut oil; Corylus avellana; fatty acids; triacylglycerol 2-position; tocopherols;
tocotrienols; triacylglycerols; sterols; steradienes; iodine value

INTRODUCTION

Hazelnuts provide a ready source of a nutritionally valuable

oil, the kernels containing 60 wt % of oil, which can be
recovered effectively by pressing. Hazelnut oil is relatively rich
in mono- and polyunsaturated fatty acids, which are believed
to have beneficial effects in the prevention of coronary heart
diseases, and regular consumption of nuts has proven to reduce
the incidence of such diseases and to increase longevity (1, 2).
About 700 000 metric tons of hazelnuts are produced worldwide
every year (3). Approximately 70% of world production is from
Turkey, with smaller quantities from the United States, Italy,
France, and Spain. Production in Turkey has increased considerably in recent years (4).
Hazelnuts can be obtained from various species of Corylus.
In Turkey, the major producing country, different varieties of
the genus Corylus are used, principally the common hazelnut
C. aVellana var. pontica and also the giant hazelnut C. aVellana
maxima Mill and the Turkish hazelnut C. colurna var.
glandulifera. Wild varieties interbreed with cultivated ones, and
different varieties are frequently grown and harvested together.
Hazelnut oils are often sold to the speciality product market
due to its distinctive flavor; the oil is therefore in demand as a
salad oil and for flavoring confectionery. The flavor of hazelnut
oil can be increased significantly by roasting the nuts prior to
pressing; thus, the hazelnut oils sold on the market can be crude
products from roasted or unroasted nuts, refined oils from
unroasted nuts, or mixtures of these.
* Author to whom correspondence should be addressed (e-mail
[email protected]; fax +44 1904 461 111).
Central Science Laboratory.
Eurofins Scientific.
# Eurofins/Wiertz-Eggert-Jorissen.

The composition of edible vegetable oils has been studied

for many years with a view to obtaining information to improve
the quality of the product in terms of storage stability, flavor,
taste, and nutrition and also to ensure the authenticity of the
material.
A number of analytical methods have been developed for
the analysis of vegetable oils, and those most frequently applied
have been tested by in-house validation or collaborative trial
and have been adopted as official procedures by bodies such
as the Association of Official Analytical Chemists, the American
Oil Chemists Society, and the International Olive Oil Council.
Many tables describing the composition of the major and
selected minor components of vegetable oils have been produced
and published in standard texts. The Codex Alimentarius
Committee on Fats and Oils has published standards for a
number of named vegetable oils, which are based on tables of
composition (5). Data for the composition of an oil sample can
be compared with the ranges published in the Codex standard
to help to judge whether the sample is authentic and of sufficient
purity for trading purposes. However, the usefulness of the
trading standards is limited by the accuracy of the data contained
in the tables, and the scientific literature occasionally provides
evidence of data from authentic oils that falls outside the Codex
ranges.
There are likely to be many reasons for differences between
genuine results and published data. The two most important are,
first, that the number of samples taken for analysis can be
insufficient to cover the variety of composition found in nature
and, second, that advances in technology continually improve
the qualitative and quantitative performance of the measuring
instruments. The latter is best illustrated in the increase in

10.1021/jf047836w CCC: $30.25 2005 American Chemical Society

Published on Web 05/13/2005

4844

J. Agric. Food Chem., Vol. 53, No. 12, 2005

resolution of chromatographic separation systems used in many

analyses over the past 40 years.
The purpose of this survey was to provide data for the
composition of hazelnut oil, which is not well represented in
terms of sample numbers in the current Codex standard. To
this end, samples of known origin were obtained from specific
oilseed-producing countries and extracted by means as similar
as possible to those used in industry. The oils were analyzed
using standardized methods incorporating modern quality assurance procedures.
A similar study carried out in 1995 reported the results of a
survey of the composition of hazelnuts (6). The survey was
limited to 10 samples of nuts collected from eight geographical
locations with the oil obtained by solvent extraction in the
laboratory. Various analyses were carried out using British
Standard and IUPAC methods and the results compared with
some values published in the literature.
For the survey reported here a larger number of samples (32)
were collected from a reduced number of countries (five) that
provided 85% of the worlds hazelnut production in 1993 (3)
and the major market share for oil production. A full audit trail
was established to ensure the authenticity of the samples. Oil
was extracted by pressing under industry conditions and
analyzed using official methods with full quality assurance
procedures in place.
To observe any effect of roasting on composition, portions
of five of the hazelnut samples were roasted in the laboratory
under conditions that simulated industry practice. These samples
were subjected to the same analyses as the oils from unroasted
nuts.
MATERIALS AND METHODS
Hazelnuts. It was not possible to confirm the identity of any nut
sample in terms of variety, and it was evident from the size and shape
of nut and the thickness of the shell that various varieties were present.
However, each batch of samples appeared to be homogeneous in terms
of shell size, shape, and color, and all nuts were in good condition on
delivery.
Hazelnuts, in batches of 4 kg each, were collected from Croatia (one
sample), France (nine samples), Italy (eight samples), Spain (eight
samples), and Turkey (six samples). Two samples, one each from France
and Croatia, were collected from the year 2000 harvest and the
remainder from 2001. To guarantee the authenticity of the oils, the
nuts were collected by the authors in the country of origin. It was
ensured as closely as possible that the grade and variety were intended
and suitable for oil production, that is, that they was not a type normally
used for direct consumption or confectionery manufacture.
The nuts were packed in secure heavy-duty polyethylene containers
and sealed with numbered tamper-proof plastic tags with a unique
number that was used at all stages from collection through analysis,
allowing a detailed audit trail to be established. They were stored at 4
C prior to removal of the kernels. The hazelnuts were shelled by hand
as there was no commercial means of shelling nuts in the United
Kingdom. The kernels were stored in sealed bags in the dark at 4 C
until transport to the oil-pressing plant. The samples and their origin
are listed in Table 1.
Roasted Hazelnuts. To observe any effect of roasting on composition, portions of five of the hazelnut samples were roasted in the
laboratory under conditions that simulated industry practice (7). Nuts
were roasted in a laboratory oven (Gallenkamp Hotbox Size 1, Sanyo
Gallenkamp, Loughborough, U.K.), which had been fitted with a
rotating shaft and drum turned by an external drive shaft at 30 rpm.
The oven was preheated to 165 C and held for 30 min. Hazelnut kernels
were introduced quickly and roasted for 25 min at 165 C. The roasted
nuts were transferred to sealed bags and allowed to cool and then stored
at 4 C prior to pressing.

Crews et al.
Table 1. Origin and Collection Date of Samples
country

region

collection date

Croatia
France
France
France
France
France
France
France
France
Italy
Italy
Italy
Italy
Italy
Italy
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Turkey
Turkey
Turkey
Turkey
Turkey
Turkey

Unje, Black Sea coast

Pillac
Farranans, Isere
Cancon
Cancon
Haute Pyrenees
Cancon
Cancon
Cancon
Sangano, Piedmont
Busca, Piedmont
Torino
Petralia
Provincia di Torino
Provincia di Cuneo
Porrera, Catalonia
Reus, Catalonia
Al Forga, Catalonia
Riudons, Catalonia
Espinaredo, Asturias
Villaviciosa, Asturias
Catalogne
Black Sea coast, Istanbula
Black Sea coast, Istanbul
Eninonu, Istanbul
Eninonu, Istanbul
Istanbul
Istanbula

August 2000
December 2000
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
August/September 2001
August 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
August 2001
August 2001
September 2001
August 2001
September 2001
August 2001
August 2001
August 2001
August 2001
August 2001
August 2002

Supplied as kernels, otherwise in the shell.

Hazelnut Oil. The oil was extracted from the hazelnut kernels by
cold pressing in an industry pilot scale screw press (Statfold Seed Oils,
Tamworth, U.K.). The batch size and processing rates were adjusted
to maintain the sample temperature at 37 C. The oils were then
filtered and stored in polypropylene bottles under nitrogen at 4 C in
the dark prior to analysis. Normal steam cleaning procedures were
carried out between samples to ensure that there was no adventitious
mixing of different samples.
Standards. Standards of the sterols stigmasterol (95%), and sitosterol
(containing campesterol) and of fatty acid methyl esters were obtained
from Sigma, Poole, U.K. The steradienes stigmasta-3,5-diene (69%)
with campesta-3,5-diene (13%) and stigmasta-3,5,22-triene (97%) were
obtained from Chiron AS, Trondheim, Norway. The internal standard
cholesta-3,5-diene (95%) was obtained from Sigma. Tocopherols and
tocotrienols (>95%) were obtained from Merk Biosciences, Schwalbach, Germany. The CBR reference sample RM162 (soy and maize
oil) was purchased from the Community Bureau of Reference, Brussels,
Belgium.
Determination of Sterols. The composition of the sterols (desmethyl
sterols) was determined according to the International Standards
Organization method (8). Gas chromatography (GC) was carried out
with a Varian model CP3380 gas chromatograph (Varian Inc., Walton,
U.K.) fitted with a JW DB-5 fused silica column (60 m 0.25 mm
i.d. 0.1 m film thickness, Fisher Scientific, Loughborough, U.K.).
Helium carrier gas was used at a pressure of 30 psi, 1 mL/min. The
injector and detector temperatures were 320 C, and the oven was
programmed from 240 to 255 C at 4 C/min. The injection volume
was 1 L, with a split ratio of 20:1.
For quality assurance purposes two reference materials were
analyzed. These were a mixture of commercial standards comprising
sitosterol, campesterol, and stigmasterol and a characterized in-house
mixture of equal volumes of sunflower oil and rapeseed oil. Batches
for GC comprised a reagent blank, the standard mixture, the reference
oil, five samples, and one sample injected in duplicate. Acceptance
criteria were that results for the samples injected in duplicate, the
standards, and the reference oils had to agree to within (10% of the
mean values.
Determination of Steradienes. The steradienes campesta-3,5-diene,
campesta-3,5,22-triene, stigmasta-3,5-diene, and stigmasta-3,5,22-triene

Main Constituents of Authentic Hazelnut Oils

and their major isomers were determined according to the European
Commission method for stigmastadienes in olive oil (9). The sensitivity
and scope of the method were increased by the incorporation of mass
spectrometric detection, the use of internal standardization, and the use
of calibration graphs as published previously (10).
GC-MS was performed using a Voyager instrument (ThermoFinnigan, Hemel Hempstead, U.K.) fitted with a JW DB-WAXETR
column (30 m 0.25 mm, Fisher Scientific) with a 0.25 m phase.
The GC used helium carrier gas at a pressure of 15 psi. The injector
and transfer line temperatures were 260 C, and the column oven
temperature was held isothermally at 250 C. Splitless injections of 1
L were made. The mass spectrometer was operated in selected ion
monitoring mode, recording the response for the molecular ions of the
steradienes and, for confirmation purposes, the sterol ring fragment at
m/z 255.
For quality assurance purposes an in-house mixture of virgin olive
oil with refined olive oil blended to contain 0.15 mg/kg stigmastadienes was analyzed as a reference material. Extraction batches
comprised a reagent blank, the in-house reference mixture, and 6-10
sample oils. Analytical (GC-MS) batches comprised a reagent blank,
standards plus one or two of the extraction batches above. Acceptance
criteria were that the levels of individual steradienes in the blank did
not exceed 0.05 mg/kg and that the stigmastadienes content of the
reference oil fell within 0.1-0.2 mg/kg.
Determination of Fatty Acids (FA). The FA composition of the
oils was determined according to the method the International Union
of Pure and Applied Chemistry method (11, 12). GC was carried out
with a Varian model CP3380 gas chromatograph (Varian Inc., Walton,
U.K.) and a fused silica column (50 m 0.35 mm i.d.), coated with
CP-Sil 88 (0.2 m film thickness, Varian Inc.). Helium carrier gas was
used at a pressure of 10 psi, giving a flow rate of 1 mL/min. The injector
and detector temperatures were 250 C, and the oven was programmed
from 160 to 220 C at 1 C/min. The injection volume was 1 L, with
a split ratio of 20:1.
For quality assurance purposes a Community Bureau of Reference
(CBR) sample, RM162 (soya and maize oil), was used. Each analytical
batch included a reagent blank, a range of standards, the CBR reference
oil, 10-14 samples, 3 samples prepared in duplicate, and 3 samples
injected in duplicate. A midrange standard was re-injected at the end
of the batch as a back-calibration to check instrument stability.
Acceptance criteria for the batch were that the FA concentrations of
the CBR and back-calibration standards had to fall between 95 and
110% of the nominal values. Results for the duplicate preparations had
to fall within 2% of the mean values for palmitic acid (C16:0), stearic
acid (C18:0), oleic acid (C18:1), and linoleic acid (C18:2) and within
10% of the mean values for myristic acid (C14:0), palmitoleic acid
(C16:1), linolenic acid (C18:3), arachidic acid (C20:0), eicosenoic acid
(C20:1), behenic acid (C22:0), and lignoceric acid (C24:0). Results
for the samples injected in duplicate had to fall within 2% of the mean
values for C16:0, C18:0, C18:1, and C18:2, within 5% of the mean
values for C16:1 and C18:3, and within 10% of the mean values for
C14:0, C20:0, C20:1, C22:0, and C24:0.
Determination of FA in the Triacylglycerol 2-Position. The
composition of FA in the triacylglycerol (TAG) 2-position was
determined according to IUPAC method 2.210 (13). GC was carried
out with a Hewlett-Packard model 5890 gas chromatograph (Agilent
Technologie, Waldbronn, Germany) and a fused silica column 60 m
0.25 mm coated with BPX70 (0.25 m film thickness, SGE,
Darmstadt Germany). Hydrogen carrier gas was used at a pressure of
21 psi, giving a flow rate of 1.2 mL/min. The injector and detector
temperatures were 270 C, and the oven was programmed from 80 to
120 C at 20 C/min, then to 200 C at 6 C/min, and then at 4 C/min
to 225 C and held for 8 min. The injection volume was 1 L. For
quality assurance purposes an in-house reference material (cocoa butter)
was analyzed. Each batch included a reagent blank, the cocoa butter
reference sample, and up to 20 samples. Two samples in each batch
were analyzed in duplicate. Acceptance criteria were that the results
for the cocoa butter reference material had to fall within 100 ( 10%
of the nominal value, and the results for the duplicate preparations had
to fall within 2% of the mean values for C16:0, C18:0, C18:1, and
C18:2.

J. Agric. Food Chem., Vol. 53, No. 12, 2005

4845

Determination of Tocopherols and Tocotrienols. The tocopherol

and tocotrienol composition was determined according to IUPAC
method 2.432 (14). The HPLC system comprised a Merck-Hitachi
model L-6000A pump (Merck, Darmstadt, Germany) and a 250 mm
4 mm LiChrospher 100 diol column (particle size ) 5 m, Merck).
The eluent was 4% tert-butylmethyl ether in hexane at a flow rate of
2.3 mL/min. The detector was a Hitachi model L-4000 UV operated
at 295 nm (Merck). Injections (50 L) were made using a Bischoff
model 728 autosampler (Bischoff Chromatography, Leonberg, Germany). Tocopherol and tocotrienol reference standards were obtained
from Calbiochem-Merck Biosciences, Schwalbach, Germany. The
reference material was a vapor condensate of mixed vegetable oils
enriched in tocopherols and tocotrienols. Each batch included a reagent
blank, a reference material, a standard mixture, and up to 20 samples.
Two samples in each batch were analyzed in duplicate. Acceptance
criteria were that results for the reference sample had to fall within
10% of the characterized value and that results for the duplicate analyses
had to agree within 5% of the mean value.
Determination of the Triacylglycerol Carbon Number (TCN).
The TCN was determined according to the AOCS method (15). GC
was performed using a Carlo Erba HRGC 5300 Mega series (CE
Instruments, Wigan, U.K.). The GC was fitted with a JW DB-5HT
column (15 m 0.32 mm with a 0.1 m film thickness; Fisher
Scientific) with a helium carrier gas pressure of 12 psi. Cooled oncolumn injections (1 L) were made with the oven programmed from
100 C (held 1 min) to 310 C at 49 C/min and then at 4 C/min to
390 C.
For quality assurance purposes a characterized in-house reference
mixture of sunflower and rapeseed oils was used. Each batch comprised
a reagent blank, a range of standards, the reference oil, 15 samples,
and 1 sample analyzed in duplicate. Acceptance criteria were that results
for the in-house reference oil and for samples injected in duplicate had
to agree to within 10% of the mean values.
Determination of the Iodine Value. The iodine value was calculated
from the FA composition according to the American Oil Chemists
Society method (16). The data were obtained from the analysis of all
samples including the averaged results for three samples prepared in
duplicate and the three injected in duplicate. Acceptance criteria were
that the iodine values of the reference standard had to fall within 100
( 10% of those calculated from the FA CBR composition data of the
CBR reference material.
Statistical Analysis. The data were examined using the CSLMetabolab visualization toolbox, a graphical user interface for the
Matlab program (The MathWorks Ltd., Cambridge, U.K.) allowing
multivariate and classical statistical analyses and methods for rapid
integration and spectral visualization. The statistical methods employed
were principal components analysis (PCA) for initial data exploration,
one-way analysis of variance (ANOVA) to detect whether one or more
oils were significantly different from the others for a given factor, and
the t test for pairwise comparison of different oil types.
RESULTS AND DISCUSSION

Oil Extraction. The volumes of oil returned from the press

varied considerably between samples, apparently as a factor of
the physical nature of the kernel. They were generally smaller
than would be expected from a full-scale oil-pressing plant. For
several of the roasted nuts the yield of oil was very low. This
was contrary to industry practice and was believed to be due to
technical factors related to the size of the press and the softness
of their kernels, and not to losses on roasting. As the quantity
of oil obtained in most cases clearly differed from the actual
volume present in the kernel, no data are presented for the oil
content of the samples.
Quality Assurance. For the data reported all results for the
replicate injection and analysis samples were within the warning
limits of Shewart control charts (17), and all results for the
reference materials were within the specified limits.
Sterol Composition. The sterol composition of hazelnut oils
expressed in milligrams per 100 g of oil is given in Table 2.

4846

J. Agric. Food Chem., Vol. 53, No. 12, 2005

Crews et al.

Table 2. Sterol Composition of Hazelnut Oils (Milligrams per 100 g)

unroasted

roasted

sample

Croatia

France

Italy

Spain

Turkey

Croatia

Italy

Spain

no.
cholesterol
cholestanol
brassicasterol
24-methylenecholesterol
campesterol
campestanol
stigmasterol
7-campesterol
5,23-stigmastadienol
clerosterol
-sitosterol
sitostanol
5-avenasterol
5,24-stigmastadienol
7-stigmastenol
7-avenasterol

1
0.7
0.7
nd
nd
10
0.6
1.6
nd
nd
1.6
134.1
4.8
4.5
0.4
2.4
1.2

8
nda0.7
nd-0.7
nd
nd
7.310.1
0.40.9
1.42.7
nd1.1
nd
1.52.4
112.5161.2
3.26.7
4.56.1
0.31.5
nd3.2
nd0.8

6
nd0.5
nd
nd
nd
6.78.1
nd0.8
2.05.4
nd1.7
nd-0.4
1.82.8
116.4138.8
3.27.6
4.26.2
nd1.4
nd2.4
nd1.8

7
nd1.4
nd
nd
nd
5.99.7
nd0.7
1.51.9
nd0.6
nd
1.94.0
92.9149.1
2.96.2
3.86.5
nd0.8
nd1.5
nd0.7

6
nd0.7
nd
nd
nd
5.810.7
nd1.3
1.52.8
nd0.6
nd
1.83.8
84.5143.1
3.311.7
0.89.4
nd2.0
nd2.1
nd2.2

1
nd
nd
nd
nd
9.0
nd
1.6
0.3
nd
1.4
122.5
4.1
4.6
1
2.5
1.2

2
0.30.6
nd
nd
nd
6.16.6
0.3
1.6
nd0.5
nd
2.72.9
110.8118.2
4.06.0
2.25.0
0.1
nd
nd1.3

3
0.21.1
nd
nd
nd
7.18.9
nd0.5
1.51.8
nd0.9
nd
2.22.7
119.3144.5
3.96.7
2.26.2
nd0.7
nd0.3
nd1.0

total

162.6

135.3195.2

141.3177.2

114.2178.4

105.7183.2

148.3

131.0140.2

139.2171.5

nd, <0.1.

Table 3. Sterol Composition of Hazelnut Oils (Percent m/m)

unroasted

roasted

sample

Croatia

France

Italy

Spain

Turkey

Croatia

Italy

Spain

no.
cholesterol
cholestanol
brassicasterol
24-methylenecholesterol
campesterol
campestanol
stigmasterol
7-campesterol
5,23-stigmastadienol
clerosterol
-sitosterol
sitostanol
5-avenasterol
5,24-stigmastadienol
7-stigmastenol
7-avenasterol

1
0.4
0.4
nd
nd
6.1
0.3
1.0
0.2
nd
1.0
82.4
3.1
2.9
0.3
1.4
0.7

8
nda0.5
nd0.4
nd
nd
4.86.3
0.20.6
0.91.8
nd0.7
nd
1.01.6
82.884.7
2.14.1
3.04.2
0.31.0
nd1.6
nd0.5

6
nd0.4
nd
nd
nd
4.65.1
nd0.5
1.13.0
nd0.9
nd0.3
1.31.9
80.184.3
2.35.2
3.04.6
nd0.6
nd1.7
nd1.3

7
nd0.8
nd
nd
nd
4.95.5
nd0.7
0.91.4
nd0.4
nd
1.32.3
81.385.9
2.45.1
2.65.0
nd0.5
nd1.2
nd0.6

6
nd0.6
nd
nd
nd
4.26.3
nd0.8
1.5
nd0.6
nd
1.42.8
78.883.3
2.44.0
2.85.1
nd1.2
nd0.6
nd1.8

1
nd
nd
nd
nd
6
nd
1
nd
nd
1
82
3
3
1
2
1

2
0.20.4
nd
nd
nd
4.35.0
0.2
1.2
nd0.4
nd
1.92.2
84.384.6
3.24.8
1.63.1
0.1
nd
nd1.0

3
0.10.8
nd
nd
nd
4.95.2
nd0.3
1.01.2
nd0.6
nd
1.31.8
84.385.7
2.44.7
1.73.3
nd0.5
nd0.2
nd0.6

nd, <0.1.

The levels in oils from unroasted nuts ranged from about 135
to 200 mg/100 g. These values are lower than those of one
earlier survey (6) but are in good agreement with the results
other surveys (18-21). The dominant sterols were -sitosterol
and campesterol. Levels of sitostanol and 5-avenasterol were
very variable, from about 3 to 12 mg/100 g and from 1 to 9
mg/100 g, respectively. The range of levels in oils from roasted
nuts fell within that for the unroasted samples.
The absolute level of sterols is not of great value in
authenticity testing compared to the relative composition, where
the variation between samples of the same type is smaller. A
considerably higher level of 5-avenasterol (21 mg/100 g) has
been reported in one sample of hazelnut oil (22); however, the
desmethyl fraction was not separated prior to GC, and the peak
identified as 5-avenasterol might have included coeluting
compounds. Others have reported similarly wide ranges of levels
of 5-avenasterol, 7-avenasterol, and 7-stigmastenol (19).
The sterol composition of hazelnut oils expressed in percent
m/m of total sterols is given in Table 3.
Values for the percentage composition of sterols were
generally in good agreement with the results published elsewhere

(18, 20, 21, 23-25). Lower campesterol and stigmasterol but

higher sitosterol levels have been reported in 20 Spanish
hazelnut oils (26), possibly as a consequence of lower chromatographic resolution.
Values for most sterols were similar to those found in the
earlier survey (6). -Sitosterol ranged from 78.8 to 85.9%
(average ) 82.7%), at the upper end of the range (74.9- 84.4%,
average ) 81.1%) reported, as were the proportions of 5,24stigmastadienol (nd-1.2%).
Similar values for cholesterol, campesterol, stigmastenol, and
sitosterol have been reported in four Italian hazelnut oils (27)
but with slightly higher levels of 7-stigmastenol (1.2-1.6%),
possibly unresolved from neighboring peaks as no other sterols
were measured.
Levels of the later eluting minor sterols were lower than most
results reported elsewhere (6). For 7-stigmastenol, 14 samples
(50%) were below the previous minimum value of 0.2%, and
for 7-avenasterol, 20 (71%) were below the minimum of 0.4%.
Higher levels of these minor sterols (5-avenasterol at 6.18.5%, 7-avenasterol at 1.9-3.3%, and 7-stigmasterol at 1.12.9%) were found in six New Zealand oils (28), and very high

J. Agric. Food Chem., Vol. 53, No. 12, 2005

Main Constituents of Authentic Hazelnut Oils

4847

Table 4. Fatty Acid Composition of Hazelnut Oils (Percent m/m)

unroasted

roasted

sample

Croatia

France

Italy

Spain

Turkey

Croatia

Italy

Spain

no.
myristic acid, C14:0
pentadecylic acid, C15:0
palmitic acid, C16:0
palmitoleic acid, C16:1
margaric acid, C17:0
heptadecenoic acid, C17:1
stearic acid, C18:0
oleic acid, C18:1
linoleic acid, C18:2
linolenic acid, C18:3
arachidic acid, C20:0
eicosenoic acid, C20:1
behenic acid, C22:0
docosenoic acid, C22:1
lignoceric acid, C24:0

1
nda
nd
5.1
0.2
nd
0.1
2.7
82.6
8.7
0.2
0.1
0.2
nd
nd
nd

8
nd0.1
nd
5.06.3
0.20.3
nd0.1
0.1
1.62.9
75.681.9
9.4 to17.1
nd0.4
0.1
0.10.2
nd
nd
nd

6
nd
nd
5.86.6
0.20.3
nd0.1
0.1
2.42.8
80.383.4
6.210.6
0.11.0
0.10.2
0.10.2
nd
nd
nd

7
nd0.1
nd
5.36.1
0.20.3
nd
0.1
1.92.5
75.381.8
10.115.9
0.10.2
0.1
nd0.2
nd0.1
nd
nd

6
nd
nd
5.16.3
0.10.3
nd
0.1
2.03.1
75.782.9
8.115.4
0.10.3
0.1
0.10.2
nd
nd
nd

1
nd
nd
5.2
0.2
nd
0.1
2.7
81.1
9.7
0.6
0.2
0.2
nd
nd
nd

2
nd
nd
6.26.6
0.20.3
0.1
0.1
2.42.6
83.383.8
6.57.2
0.1
0.1
0.1
nd
nd
nd

4
nd
nd
5.26.2
0.2
nd
n-0.1
2.12.5
78.582.3
9.812.0
0.10.5
0.1
0.10.2
nd
nd
nd

nd, <0.1.

Table 5. Iodine Values of Hazelnut Oils

unroasted

roasted

sample

Croatia

France

Italy

Spain

Turkey

Croatia

Italy

Spain

no.
iodine value

1
87.0

8
87.195.4

6
84.788.1

7
88.093.0

6
85.891.9

1
88.5

2
83.984.7

4
88.289.8

proportions of 5-avenasterol were found in a single Spanish

sample (22).
There was no evidence of a significant effect of roasting on
the desmethylsterol composition, with the range of sterols falling
within the ranges of the oils from unroasted nuts except for
5-avenasterol, which was reduced in the roasted samples.
Steradiene Composition. No steradienes (stigmastadiene,
campestadiene, stigmastatriene, and campestatriene) were detected in any hazelnut oil before or after roasting, subject to a
limit of detection of 0.02 mg/kg. There were no reports of the
analysis of hazelnut oils for sterol degradation products in the
literature. Steradienes are formed by dehydration of sterols,
notably during refining, and would not be expected to be formed
in oils that had not undergone severe treatment in processing.
FA Composition. The FA composition of the hazelnut oils
is given in Table 4. The values obtained were in very close
agreement with the ranges reported in the literature (4, 6, 1823, 25, 27, 29-36).
One sample, from Italian nuts, had a C18:3 value of 1.0%,
which was notably higher than the other samples. The C18:0
value for another sample from French nuts (1.6%) was lower
than for any other sample, but higher than a single oil of 17
Spanish oils, which contained 1.4% (18). The C18:0 value for
one sample from Turkish nuts was high (3.1%) but within the
range reported elsewhere (18).
Some workers have reported higher values of C20:0 and
C20:1 than reported here (34, 37). In oil from 22 Spanish
(Catalonia) nuts of six varieties C20:0 was found at 0.2-0.3%,
twice the level reported here (38). Various other workers have
published FA values similar to those reported here but with the
proportions of some acids at slight variance with our results.
Higher proportions of C18:3 were measured in eight Italian oils
(39). Lower C18:0 and C18:1 and higher C18:2 and C18:3 levels
were measured in 17 varieties in the United States (40), and
lower C18:0 levels were measured in 6 oils from New Zealand
(28).

Results for the oils from Turkish hazelnuts were in general

agreement with other results for 10 oils from that country (41).
Lower levels of C16:0 (4.6-5.3%) were reported from the
analysis of the oils from 10 German varieties using a packedcolumn GC (42).
No marked changes were observed in the FA composition
of oils from roasted nuts.
Iodine Value. The mean calculated iodine values of hazelnut
oils are given in Table 5.
Three samples, two from French nuts and one from Spanish
nuts, had iodine values slightly above the maximum of 92.8
reported earlier (6) at 93.1, 95.4, and 93.0, respectively;
however, the spread of the results was comparable. Slightly
lower values have been reported for eight Italian cultivars (39).
The iodine values of oils from roasted nuts fell within the range
of the oils from unroasted nuts.
FA in the Triacylglycerol 2-Position Composition. The
triacylglycerol 2-position fatty acid composition of hazelnut oils
is given in Table 6.
One Italian oil had very low C16:0 (0.2%) and very high
C18:3 (1.0%). The high value for C18:3 was mirrored by that
of the total FA composition result, but that for C16:0 was not.
For the oils from roasted nuts the values fell within those of
the unroasted samples with the notable exception of the Italian
oil described above, which surprisingly had very high C16:0
(4.3%) and 18:0 (1.9%) and very low C18:2 (7.7%).
TCN. The TCN composition of the hazelnut oils is given in
Table 7.
The results were similar to those reported elsewhere (6, 19)
except that for C56 was detected in a greater proportion of the
samples, and two samples from Italian nuts and one each from
Spanish and Turkish nuts had higher levels of C50.
There are few other reported investigations into the triacylglycerol profiles of hazelnut oils, although measurement of
trilinolein is of relevance in detecting the presence of seed oils
in olive oil. Published papers (24, 43-45) have reported the

4848

J. Agric. Food Chem., Vol. 53, No. 12, 2005

Crews et al.

Table 6. Triacylglycerol 2-Position Fatty Acid Composition of Hazelnut Oils (Percent m/m)
unroasted

roasted

sample

Croatia

France

Italy

Spain

Turkey

Croatia

Italy

Spain

no.
myristic acid, C14:0
palmitic acid, C16:0
palmitoleic acid, C16:1
margaric acid, C17:0
heptadecenoic acid, C17:1
stearic acid, C18:0
oleic acid, C18:1
linoleic acid, C18:2
linolenic acid, C18:3
arachidic acid, C20:0
eicosenoic acid, C20:1
others

1
nda
0.7
0.2
0.1
nd
0.3
86.1
12.5
0.1
nd
nd
nd

8
nd
0.20.8
0.10.2
0.1
nd
0.10.4
75.685.3
13.923.6
0.10.3
nd
nd
nd0.4

6
nd
nd0.3
0.10.2
0.1
nd
0.10.2
84.389.6
9.014.9
0.11.0
nd
nd
nd0.3

7
nd
0.30.7
0.10.2
0.1
nd
0.10.3
76.584.5
14.522.1
0.10.2
nd
nd
nd0.1

6
nd
0.31.3
0.1
0.1
nd
0.20.7
78.788.2
11.020.0
0.10.2
nd
nd
nd

1
nd
0.4
0.1
0.1
nd
0.2
84.3
14.4
0.6
nd
nd
0.1

2
nd
0.30.4
0.10.2
0.1
nd
0.21.9
85.888.4
7.710.7
nd0.1
nd
nd0.1
nd0.4

3
nd
0.30.5
0.1
0.1
nd
0.20.3
80.284.0
15.319.0
0.10.4
nd
nd
nd0.4

nd, <0.1.

Table 7. Triacylglycerol Carbon Number (TCN) Composition of Hazelnut Oils (Percent m/m)
unroasted

roasted

sample

Croatia

France

Italy

Spain

Turkey

Croatia

Italy

Spain

no.
TCN 48
TCN 50
TCN 52
TCN 54
TCN 56
TCN 58
TCN 60

1
nda
0.8
14.9
81.7
1.2
0.1
nd

8
nd
0.41.1
13.818.5
78.983.8
nd0.9
nd0.1
nd

6
nd0.1
0.81.8
15.519.4
77.081.3
nd1.4
nd
nd

7
nd
0.71.6
14.516.9
79.982.9
nd1.0
nd
nd

6
nd
0.71.8
14.417.5
79.483.3
nd0.5
nd
nd

1
nd
0.9
17.3
80.8
1
nd
nd

2
nd0.1
0.91.0
17.518.8
79.280.4
11.1
nd0.4
nd

3
nd0.3
0.91.1
15.717.5
78.982.1
1.01.81
nd0.4
nd

nd, <0.1.

Table 8. Tocopherol and Tocotrienol Composition of Hazelnut Oils (Milligrams per Kilogram)
unroasted
sample

Croatia

France

no.
R-tocopherol
-tocopherol
-tocopherol
-tocopherol
R-tocotrienol
-tocotrienol
-tocotrienol
-tocotrienol

1
340
nda
65
nd
nd
nd
nd
nd

8
343545
nd10
nd10
nd10
nd
nd
nd
nd

total tocols

405

343545

Italy

roasted
Spain

Turkey

Croatia

Italy

Spain

6
118440
nd10
nd10
nd10
nd209
nd
nd
nd

7
314590
nd15
nd42
nd10
nd
nd-22
nd-34
nd

6
319492
nd10
nd130
nd
nd
nd
nd
nd

1
340
77
nd
10
nd
nd
nd
nd

2
346362
nd11
nd21
nd
nd
nd
nd
nd

3
368548
nd12
nd52
nd
nd
nd
nd
nd

354500

314613

351554

427

367

372610

nd, <10.

triacylglycerol profiles as separated by triacylglycerol rather than

TCN, and the coelution in the published profiles of some
triacylglycerols with and without C16:0 acid means that
conversion of published data to TCN values was not possible.
The C54 proportions measured appeared to decrease in all
but one sample on roasting, but these changes were not
statistically significant when compared with the standard
deviation of results for the reference oil.
Tocopherol Composition. The tocopherol composition of
hazelnut oils is given in Table 8.
The major tocopherol present in all samples was R-tocopherol,
the range of values for which was large, with one Italian sample
having a relatively very low content (118 mg/kg). A broadly
similar range (211-655 mg/kg) has been reported elsewhere
for 17 hazelnut oils, and this also contained a similarly atypically
low level (94 mg/kg) in one oil (18).
Levels of R-tocopherol were in good agreement with results
published elsewhere (19, 23, 41, 46-49). Oils from six nuts

grown in New Zealand were shown to have a wider range of

R-tocopherol and contained -, -, and -tocopherols (28).
Slightly lower levels of R-tocopherol were present in three
samples grown in the United States, with a much lower level
in a sample from the wild species C. cornuta (35). In fact,
several workers have reported higher incidences of and -tocopherol than measured in this study (23, 40, 46,
50, 51). The possibility of using - to -tocopherol ratios
to detect hazelnut oil addition to olive oil has been suggested
(52).
The wide range of -tocopherol, absent in several samples
but present at up to 130 mg/kg, and the low incidence of samples
with -tocopherol or any of the tocotrienols to some extent
reflected published values.
Tocotrienols were detected in two samples. A sample from
Italian nuts contained R-tocotrienol as its principal tocol at
209 mg/kg, and one from Spanish nuts contained - and
-tocotrienols at 22 and 34 mg/kg, respectively. No tocotrienols

Main Constituents of Authentic Hazelnut Oils

Figure 1. ANOVA of values for C18:1 fatty acids.

were reported in a limited number of samples studied earlier

(6, 46).
Erratic changes were found in the levels of R- and -tocopherols when hazelnuts were roasted. There was also an increase
in measured -tocopherol in three of the six roasted samples.
However, the significance of these changes is difficult to judge.
Loss of vitamin E, mostly R-tocopherol, have been reported to
occur with increasing roasting temperature over the range 180300 C (53).
Country of Origin. The composition of the Turkish samples
agreed with published results for 10 varieties measured in terms
of FA profiles but was lower in tocopherol content (41). The
Spanish samples were lower in C18:0 and C20:0 FA than were
two sets of nine cultivars analyzed earlier (27). Slightly higher
proportions of -sitosterol and 7-stigmasterol were present in
one of the earlier sets; however, those values were probably
enhanced by the small number of sterols (five) measured and
by coelution of some sterols.
The Italian samples agreed well in FA composition with
results for three samples analyzed in 1979 (54); again, peak
coelution was probably responsible for higher -sitosterol levels
reported in the earlier work.
Although the number of samples analyzed was greater than
that in most previous work, it remained limited in terms of the
application of statistical treatment. However, the data were
examined using the CSL-Metabolab visualization toolbox, which
allowed multivariate and classical statistical analyses and rapid
integration and spectral visualization.
In most cases the values within groups based on country of
origin appeared to be normally distributed. However, there were
certain cases when standard deviations differed between the
countries, which violates the assumptions of ANOVA and t tests.
Such tests can still provide useful indicators of countries of
origin where there are strong differences between countries, but
the p values do not have their usual statistical interpretations
regarding significance levels.
The main focus was on the search for cases where there was
some separation of ranges of results by country. ANOVA
showed that most samples from Italy were separated from most
samples from the other countries using data for the C18:1 FA
(Figure 1), the 2-position FA (e.g., C18:2, Figure 2), and the
triacylglycerol carbon number 56 (Figure 3). The latter data
showed separation of Italian from Turkish oils but was less
successful with the other countries.
Univariate statistics were performed using an ANOVA on
all data looking for cases where at least one country had a
significant difference and using pairwise t tests for all pairs of
countries and all data looking for differences between individual

J. Agric. Food Chem., Vol. 53, No. 12, 2005

4849

Figure 2. ANOVA of values for 2-position C18:2.

Figure 3. ANOVA of values for triacylglycerol carbon no. 56.

pairs of countries. The most significant comparisons (>99%)

based on ANOVA p values were for 2-position FA, tocopherols,
and CN 56. Although there were overlapping data points, most
of the oils from Turkish nuts could be separated.
A PCA was performed using all variables. The data were
both mean centered and scaled to unit variance, and 10 PC
scores were calculated. There was reasonable but imperfect
separation between oils from Italian nuts and those from Spain
and France in PC scores 3 and 5.
Factors Affecting the Composition of Hazelnut Oil. The
composition of vegetable oils, including hazelnut oil, is affected
by variety, geographical location, climatic effects, and treatment,
for example, addition of fertilizer, during growth as described
below. The samples for this study were collected from sources
used for commercial production, and taking into account the
environment and treatments that the growing trees were exposed
to was beyond the scope of the work.
The varietal effects on hazelnut oil composition have been
studied with a view to providing commerce with nuts and oils
rich in antioxidant components giving stability and health
benefits (41). Significant differences in the tocopherol and sterol
composition of oil from Spanish varieties of hazelnut grown in
the United States have been reported (18). Higher levels of
R-tocopherol were found in a New Zealand cultivar than in three
U.S. and two European cultivars grown in New Zealand (28),
but there were no significant differences in the FA or sterols.
No significant differences was found in the R-tocopherol
contents of 11 samples from 8 varieties of nuts grown in four
regions of Turkey (49). It has been concluded (33) from
measurements of single samples of 12 cultivars and varieties
that they could be distinguished by virtue of their C18:2 acid
levels, which varied to a greater extent (6.7-13.6%) than did
the other acids.

4850

J. Agric. Food Chem., Vol. 53, No. 12, 2005

Linoleic acid levels are increased by irrigation and have been

shown to vary with hazelnut variety and to different degrees
within single varieties (38). Levels of C18:1 and C18:2 in
particular differed between varieties of hazelnut originating in
Spain and between these and an Italian variety grown in Spain
(44). Similar differences were found between Spanish varieties
grown in the United States (18).
Changes in the oil composition as the nut develops have been
documented (45, 52, 55, 56). Levels of R-tocopherol increase
with kernel development with a concomitant decrease in the
levels of - and -tocopherols, whereas -tocopherol levels
increase during the early stages of kernel development and then
decrease. FA composition changes radically during kernel
development, with C14:0, C16:0, C18:0, and C18:3 decreasing
and C18:1 and C18:2 increasing over 3 months (52). Treatments
such as irrigation and the application of fertilizer also affect
the FA composition (36). The FA and tocopherol composition
of hazelnut oil changes considerably with season and with stage
of growth (55); however, as hazelnuts are harvested after falling
from the tree, the samples in this survey may be assumed to be
of similar maturity and to have been collected in the same
season.
In a comparison of four varieties of hazelnut grown in two
regions and sampled annually over three years (23) it was found
that there were significant differences between years for almost
all of the FA and significant differences between locations for
C18:0, C18:1, and C18:2. The R-tocopherol content changed
with harvest year but not variety (46). The triacylglycerol content
of the same samples did not change with variety (43).
Handling of the nuts and oils can be assumed to affect their
composition. The times between harvest of the nuts and their
collection by analysts are in most cases not available. However,
following heavy harvests in regions of high productivity such
as Turkey, hazelnuts could be stored for up to a year before
use. There are few published reports of changes of hazelnut oil
composition on storage of nuts (4, 28, 30, 37-38, 53, 58-59);
however, it has been reported (57) that tocopherols had good
stability in vegetable oils at 4 C, although other authors found
R-tocopherol losses of 50% on storage for 36 weeks at 25 C
(35). Significant decreases (10%) have been found in the levels
of C18:3 in hazelnuts stored for 12 months (58). Lesser
decreases were found for C16:0 and C18:0 and increases (2%)
in the levels of C18:2. The changes were similar regardless of
the storage temperature (3-6 C or 18-25 C) and were slightly
retarded by the use of a nitrogen atmosphere.
Awareness of the effects of factors such as the levels of proand antioxidant compounds and the saturated/unsaturated fatty
acid ratios on the stability of harvested commercial hazelnuts
has led to the breeding and testing of new varieties (41). Thus,
it is most likely that the composition of hazelnut oils in
commercial production will change as varieties having beneficial
characteristics are introduced.
It is apparent from reviews of the literature that some changes
in the pattern of lipid profiles were likely to be due to factors
associated with the equipment and procedures used in different
laboratories. Various approaches have been used for sampling,
with sizes ranging from single kernels to kilograms. Samples
have frequently been obtained from experimental plots, which
do not necessarily represent commercial systems. Within the
laboratory many techniques have been applied to the extraction
and analysis of nut oils, ranging from pressing to extraction
with organic solvents or supercritical fluids. Advances in
separation technology have had a dramatic effect, allowing the
resolution of chromatographic peaks previously reported as a

Crews et al.
single component into two or more constituents. These advances
are particularly relevant in the determination of sterol and FA
composition.
In this study attempts were made to follow the processing
used in industry up to but not including refining steps such as
washing, neutralization, winterization bleaching, and deodorization, some of which have a strong effect on the oil composition.
This paper adds to and updates the available compositional
tables for hazelnut oils. The analyses have been performed using
modern equipment and validated methods incorporating stringent
quality control measures based on the use of fully characterized
reference materials. The geographical origin of the samples has
been confirmed by collection from the harvest site and the
authenticity of the nuts and oil by the use of an audit trail.
Industry conditions for the oil production were followed as
closely as possible. The data obtained by these means will be
the most relevant so far available for authenticity of traded
hazelnut oils.
LITERATURE CITED
(1) Sabate, J. Nut Consumption, Vegetarian Diets, Ischemic Heart
Disease Risk, and All-Cause Mortality: Evidence From Epidemiologic Studies. Am. J. Clin. Nutr. 1999, 70, 500S-503S.
(2) Fraser, G. E.; Sabate, J.; Beeson, W. L.; Strahan, T. M. A
Possible Protective Effect of Nut Consumption on Risk of
Coronary Heart-DiseasesThe Adventist Health Study. Arch.
Intern. Med. 1992, 152, 1416-1424.
(3) Food and Agriculture Organization. FAOSTAT data; Rome, Italy,
2004.
(4) Alasalvar, C.; Shahidi, F.; Ohshima, T.; Wanasundara, U.;
Yurttas, H. C.; Liyanapathirana, C. M.; Rodrigues, F. B. Turkish
Tombul Hazelnut (Corylus aVellana L.). 2. Lipid Characteristics
and Oxidative Stability. J. Agric. Food Chem. 2003, 51, 37973805.
(5) Codex Alimentarius; Codex Standard for Named Vegetable Oils;
Codex STAN 210-1999, 2001.
(6) Griffith, R. E.; Farmer, M. R.; Rossell, J. B. Authenticity of
Vegetable Oils Part XXIV. High-Oleic Oil SamplessHazelnut,
Apricot, Almond, High-Oleic Sunflower and High-Oleic Safflower Seed Oil. Leatherhead Food Research Association
Research ReportsAuthenticity of Edible Vegetable Oils and Fats;
Leatherhead, U.K., 1995.
(7) Saklar, S.; Katnas, S.; Ungan, S. Determination of Optimum
Roasting Conditions. Int. J. Food Sci. Technol. 2001, 36, 271281.
(8) International Standards Organization. Animal and Vegetable Fats
and OilssDetermination of Composition of the Sterol Fractions
Method Using Gas Chromatography; ISO: Geneva, Switzerland,
1999.
(9) European Commission. Characteristics of Olive Oil and OliveResidue Oil and on the Relevant Methods of Analysis; Commission Regulation EEC/2568/91 as amended. Off. J. Eur.
Communities 1991, L248, 1-83.
(10) Crews, C.; Calvet-Sarret, R.; Brereton, P. Determination of
Vegetable Fats in Chocolate by Analysis of Sterol Degradation
Products. J. Am. Oil Chem. Soc. 1997, 74, 1273-1280.
(11) International Union of Pure and Applied Chemistry (IUPAC).
Standard Methods for the Analysis of Oils, Fats and DeriVatiVes: Method 2.301. Preparation of the Fatty Acids Methyl
Esters, 1st suppl., 7th ed.; Pergamon Press: Oxford, U.K.,
1992.
(12) International Union of Pure and Applied Chemistry (IUPAC).
Standard Methods for the Analysis of Oils, Fats and DeriVatiVes: Method 2.302. Gas-Liquid Chromatography of Fatty
Acids Methyl Esters, 1st suppl., 7th ed.; Pergamon Press: Oxford,
U.K., 1992.

Main Constituents of Authentic Hazelnut Oils

(13) International Union of Pure and Applied Chemistry (IUPAC).
Standard Methods for the Analysis of Oils, Fats and DeriVatiVes: Method 2.210. Determination of Fatty Acids in the
2-Position of the Triglycerides of Oils and Fats, 1st suppl., 7th
ed.; Pergamon Press: Oxford, U.K., 1992.
(14) International Union of Pure and Applied Chemistry (IUPAC).
Standard Methods for the Analysis of Oils, Fats and DeriVatiVes: Method 2.432. Identification and Determination of Tocopherol, 1st suppl., 7th ed.; Pergamon Press: Oxford, U.K.,
1992.
(15) American Oil Chemists Society Method Ce5-86. Triglycerides
by Gas Chromatography; AOCS: Champaign, IL, 1997.
(16) American Oil Chemists Society Recommended Practice, Cd 1b87, Calculated Iodine Value; AOCS: Champaign, IL, 1997.
(17) Grant, E. L.; Leavenworth, R. S. Statistical Quality Control;
McGraw-Hill: New York, 1998.
(18) Parcerisa, J.; Richardson, D. G.; Rafecas, M.; Codony, R.;
Boatella, J. Fatty Acid, Tocopherol and Sterol Content of Some
Hazelnut Varieties (Corylus aVellana L.) Harvested in Oregon
(USA). J. Chromatogr. A 1998, 805, 259-268.
(19) Benitez-Sanchez, P. L.; Leon-Camacho, M.; Aparicio, R. A.
Comprehensive Study of Hazelnut Oil Composition with Comparisons to Other Vegetable Oils, Particularly Olive Oil. Eur.
Food Res. Technol. 2003, 218, 13-19.
(20) Bada, J. C.; Manuel, L. C.; Prieto, M.; Alonso, L. Characterization of Oils of Hazelnuts From Asturias, Spain. Eur. J. Lipid
Sci. Technol. 2004, 106, 294-300.
(21) Maguire, L. S.; OSullivan, S. M.; Galvin, K.; OConnor, T. P.;
OBrien, N. M. Fatty Acid Profile, Tocopherol, Squalene and
Phytosterol Content of Walnuts, Almonds, Peanuts, Hazelnuts
and the Macadamia Nut. Int. J. Food Sci. Nutr. 2004, 55, 171178.
(22) Parcerisa, J.; Casals, I.; Boatella, J.; Codony, R.; Rafecas, M.
Analysis of Olive and Hazelnut Oil Mixtures by HighPerformance Liquid Chromatography-Atmospheric Pressure
Chemical Ionisation Mass Spectrometry of Triacylglycerols and
Gas-Liquid Chromatography of Non-Saponifiable Compounds
(Tocopherols and Sterols). J. Chromatogr. A 2000, 881, 149158.
(23) Parcerisa, J.; Boatella, J.; Codony, R.; Farra`n, A.; Garca, J.;
Lopez, A.; Rafecas, M.; Romero, A. Influence of Variety and
Geographical Origin on the Lipid Fraction of Hazelnuts (Corylus
aVellana L) from Spain. 1. Fatty-Acid Composition. Food Chem.
1993, 48, 411-414.
(24) Gigliotti, C.; Daghetta, A.; Sidoli, A. Study of Triglyceride
Composition of Seed Oils with High Oleic Acid Content. RiV.
Ital. Sostanze Grasse 1993, 70, 533-539.
(25) Ollivier, D.; Bruckert, B.; Noyer, C.; Guerere, M.; Artaud, J.
Multicriteria Analysis for the Research of Adulteration of Virgin
Olive Oil by Hazelnut and Almond Oils. Ann. Falsif. Expert.
Chim. Toxicol. 1999, 92, 163-178.
(26) Garca-Olmedo, R.; Carballido, A.; Daz Marquina, A. Study of
the Unsaponifiable Constituents of the Spanish Nut Oils 1:
Sterolic Fraction. Anal. Bromatol. 1978, 30, 63-89.
(27) Gargano, A.; Magro, A.; Manzo, P. Chemical Characteristics of
the Fruits of Some Principal Hazelnut Cultivars 2. Ind. Aliment.
1982, 21, 15-16.
(28) Savage, G. P.; McNeil, D. L.; Dutta, P. C. Lipid Composition
and Oxidative Stability of Oils in Hazelnuts (Corylus aVellana
L) Grown in New Zealand. J. Am. Oil Chem. Soc. 1997, 74,
755-759.
(29) Gargano, A.; Magro, A.; Manzo, P. Chemical Characteristics of
the Hazelnut of Some Cultivar. Ind. Aliment. 1981, 20, 104106.
(30) Gattuso, A. M.; Indovina, M. C.; Arcoleo, G. Changes in
Chemical Composition of Kernel Lipids in Nine Hazelnut
Cultivars During Storage. Acta Hortic. 1994, 351, 649-656.
(31) Contini, M.; De Santis, D.; Anelli, G. Distribution of Fatty Acids
in Glycerides and in Free Acid Fraction of Hazelnut Oil (Corylus
aVellana L.). RiV. Ital. Sostanze Grasse 2004, 68, 405-411.

J. Agric. Food Chem., Vol. 53, No. 12, 2005

4851

(32) Botta, R.; Gianotti, C.; Richardson, D.; Suwanagul, A.; Sanz,
C. L. Hazelnut Variety Organic Acids, Sugars and Total Lipid
Fatty Acids. Acta Hortic. 1994, 351, 693-699.
(33) Botta, R.; Gianotti, C.; Me, G. Kernel Quality in Hazelnut
Cultivars and Selections Analysed For Sugars, Lipids and Fatty
Acid Composition. Acta Hortic. 1997, 445, 319-326.
(34) Mannina, L.; Patumi, M.; Fiordiponti, P.; Emanuele, M. C.;
Segre, A. L. Olive and Hazelnut Oils: A Study By High-Field
H-1 NMR and Gas Chromatography. Ital. J. Food Sci. 1999,
11, 139-149.
(35) Pershern, A. S.; Breene, W. M.; Lulai, E. C. Analysis of Factors
Influencing Lipid Oxidation in Hazelnuts (Corylus Spp.). J. Food
Process. PreserV. 1995, 19, 9-26.
(36) Beuchat, L. R.; Worthington, R. E. Technical Note: Fatty Acid
Composition of Tree Nut Oils. J. Food Technol. 1978, 13, 355358.
(37) Serra Bonveh, J.; Ventura Coll, F. Study on the Quality and
Oil Stability in the Hazelnuts Evaluation. Ann. Falsif. Expert.
Chim. Toxicol. 1997, 90, 21-28.
(38) Serra Bonveh, J.; Ventura Coll, F. Oil Content, Stability and
Fatty-Acid Composition of the Main Varieties of Catalonian
Hazelnuts (Corylus aVellana L). Food Chem. 1993, 48, 237241.
(39) Arcoleo, G. Characteristics and Composition of Some Sicilian
Hazel Nut Cultivars. RiV. Ital. Sostanze Grasse 1991, 68, 257260.
(40) Ebrahem, K. S.; Richardson, D. G.; Tetley, R. M.; Mehlenbacher,
S. A. Oil Content, Fatty Acid Composition, and Vitamin E
Concentration of 17 Hazelnut Varieties, Compared to Other
Types of Nuts and Oil Seeds. Acta Hortic. 1994, 351, 685692.
(41) Ozdemir, M.; Ackurt, F.; Kaplan, M.; Yildiz, M.; Loker, M.;
Gurcan, T.; Biringen, G.; Okay, A.; Seyhan, F. G. Evaluation
of New Turkish Hybrid Hazelnut (Corylus aVellana L.)
Varieties: Fatty Acid Composition, R-Tocopherol Content,
Mineral Composition and Stability. Food Chem. 2001, 73, 411415.
(42) Garcia, J. M.; Agar, I. T.; Streif, J. Lipid Characteristics of
Kernels from Different Hazelnut Varieties. Turkish J. Agric.
1994, 18, 199-202.
(43) Parcerisa, J.; Rafecas, M.; Casellote, A. I.; Codony, R. F. A.;
Garca, J.; Lopez, A.; Romero, A.; Boatella, J. Influence of
Variety and Geographical Origin on the Lipid Fraction of
Hazelnuts (Corylus aVellana L) from Spain. 2. Triglyceride
Composition. Food Chem. 1994, 50, 245-249.
(44) Parcerisa, J.; Boatella, J.; Codony, R.; Rafecas, M.; Castellote,
A. I.; Garca, J. L. A.; Romero, A. Comparison of Fatty-Acid
and Triacylglycerol Compositions of Different Hazelnut Varieties
(Corylus aVellana L) Cultivated in Catalonia (Spain). J. Agric.
Food Chem. 1995, 43, 13-16.
(45) Parcerisa, J.; Codony, R.; Boatella, J.; Rafecas, M. Triacylglycerol and Phospholipid Composition of Hazelnut (Corylus
aVellana L.) Lipid Fraction during Fruit Development. J. Agric.
Food Chem. 1999, 47, 1410-1415.
(46) Parcerisa, J.; Rafecas, M.; Casellote, A. I.; Codony, R.; Farra`n,
A.; Garca, J.; Gonzalez, C.; Lopez, A.; Romero, A.; Boatella,
J. Influence of Variety and Geographical Origin on the Lipid
Fraction of Hazelnuts (Corylus aVellana L.) from Spain. III: Oil
Stability, Tocopherol Content and Some Mineral Contents (Mn,
Fe, Cu). Food Chem. 1995, 53, 71-74.
(47) Dionisi, F.; Prodolliet, J.; Tagliaferri, E. Assessment of Olive
Oil Adulteration by Reversed-Phase High-Performance Liquid
Chromatography Amperometric Detection of Tocopherols and
Tocotrienols. J. Am. Oil Chem. Soc. 1995, 72, 1505-1511.
(48) Alasalvar, C.; Shahidi, F.; Liyanapathirana, C. M.; Ohshima, T.
Turkish Tombul Hazelnut (Corylus aVellana l.). 1. Compositional
Characteristics. J. Agric. Food Chem. 2003, 51, 3790-3796.
(49) Ac kurt, F.; O
zdemir, M.; Biringen, G.; Loker, M. Effect of
Geographical Origin on Vitamin and Mineral Composition of
Hazelnut (Corylus aVellena L.) Varieties Cultivated in Turkey.
Food Chem. 1999, 65, 309-313.

4852

J. Agric. Food Chem., Vol. 53, No. 12, 2005

(50) Gertz, C.; Herrmann, K. Analysis of Tocopherols and Tocotrienols in Foods. Z. Lebensm. Unters. Forsch. 1982, 14, 390394.
(51) Von Coors, U. Utilisation of Tocopherol Pattern For Recognition
of Oils and Fats Adulterations. Fett. Wiss. Technol. 1991 93,
519-526.
(52) Morchio, G.; Pellegrino, A.; Mariani, C.; Bellan, G. Feasibility
to Check the Presence of Hazelnut Oil in Olive Oil. Note 2.
RiV. Ital. Sostanze Grasse 1999, 76, 115-127.
(53) Ebrahem, K. S.; Richardson, D. G.; Tetley, R. M. Effects of
Storage Temperature, Kernel Intactness, and Roasting Temperature on Vitamin E, Fatty Acids and Peroxide Value of
Hazelnuts. Acta Hortic. 1994, 351, 677-684.
(54) Colombini, M.; Vanoni, M. C.; Amelotti, G. The Sterol
Composition of Walnut, Hazelnut, Almond and Avocado Oils.
RiV. Ital. Sostanze Grasse 1979, 10, 392-393.
(55) Ebrahem, K. S.; Richardson, D. G.; Tetley, R. M. Changes in
Oil Content, Fatty Acid Composition, and Vitamin E Composition in Developing Hazelnut Kernels. Acta Hortic. 1994, 351,
669-676.

Crews et al.
(56) Koyuncu, M. A.; Koyuncu, F.; Bostan, S. Z.; Islam, A. Change
of Fat Content and Fatty Acid Composition during the Fruit
Development Period in the Hazelnuts Tombol and Palaz Cultivars
Grown in Ordu. Acta Hortic. 1997, 445, 229-233.
(57) Von Coors, U.; Montag, A. Stability of Tocopherols in Vegetable
Oils. Fett. Wiss. Technol. 1988, 90, 129-137.
(58) Kerne, T.; Messerli, M.; Shejbal, J.; Vitali, F. The Storage of
Hazelnuts at Room Temperature Under Nitrogen (1). ReV. Choc.
Conf. Bakery 1983, 8, 24-28.
(59) Hadorn, H.; Keme, T.; Kleinert, J.; Messerli, M.; Zurcher, K.
The Behaviour of Hazelnuts Under Different Storage Conditions.
ReV. Choc. Conf. Bakery 1977, 2, 25-39.
Received for review December 22, 2004. Revised manuscript received
April 4, 2005. Accepted April 9, 2005. We thank the U.K. Food
Standards Agency for funding this project.

JF047836W