Food and Chemical Toxicology 49 (2011) 2429–2432

Contents lists available at ScienceDirect

Food and Chemical Toxicology

journal homepage: www.elsevier.com/locate/foodchemtox

Assessing the effects of gamma irradiation and storage time in energetic

value and in major individual nutrients of chestnuts
Ângela Fernandes a, João C.M. Barreira a,b, Amilcar L. Antonio a,c,d, Albino Bento a,
M. Luisa Botelho c, Isabel C.F.R. Ferreira a,⇑
a
CIMO/Escola Superior Agrária, Instituto Politécnico de Bragança, Apartado 1172, 5301-855 Bragança, Portugal
b
REQUIMTE/Departamento de Ciências Químicas, Faculdade de Farmácia da Universidade do Porto, Rua Aníbal Cunha, 164, 4099-030 Porto, Portugal
c
GTRPP/Unidade de Física e Aceleradores, Instituto Tecnológico e Nuclear, Estrada Nacional 10, 2686-953 Sacavém, Portugal
d
Departamento de Física Fundamental, Universidade de Salamanca, Plaza de la Merced, 37008 Salamanca, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Chestnut (Castanea sativa Miller) is an important food resource all over the world. In the present study, it
Received 25 May 2011 is intended to evaluate if the application of gamma irradiation doses 63 kGy maintain chestnuts chemical
Accepted 21 June 2011 and nutritional profiles unaffected. Furthermore, possible interactions among irradiation dose and stor-
Available online 28 June 2011
age time were accessed using linear discriminate analysis (LDA). The nutritional composition was eval-
uated through determination of proteins, fat, ash, carbohydrates and energetic value. The chemical
Keywords: composition was focused in the main nutrients found in chestnuts: sugars – sucrose, fatty acids – pal-
Irradiated chestnuts
mitic, oleic, linoleic and linolenic acids, tocopherols – c-tocopherol. The obtained results seem to indicate
Gamma irradiation
Energetic value
that the irradiation treatment did not affect the nutritional and chemical quality of chestnut fruits. Other-
Nutritional parameters wise, storage time exerted more evident influence in those parameters. The application of gamma irradi-
Linear discriminant analysis ation emerges as a promising technology for chestnuts chemical quality, but food safety issues have to be
evaluated in order to recommend its application as a useful conservation alternative.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction efficiency (UNEP, 2006). The immersion in cold or hot water affects
the chestnut chemical composition and may induce the develop-
Chestnut (Castanea sativa Miller) is an important food resource ment of moulds (Jermini et al., 2006; UNEP, 2006).
in several countries. In Europe, chestnut is regaining interest, with The application of gamma irradiation seems to be a promising
an increase in production area from 81,511 (2005) to 87,521 ha technology since it may achieve various effects (depending on
(2008). Portugal is among the major producers, with annual values the absorbed radiation dose) like reduced storage losses, extended
of over 20,000 t (FAOSTAT, 2010). The Trás-os-Montes region con- shelf life and/or improved microbiological and parasitological
centrates over 75% of all Portuguese production, being chestnut safety of foods. This technology had already been applied to the
one of the most important economic resources (Portuguese main commodities such as tuber and bulb crops, stored grains,
Agricultural Statistics, 2009). dried ingredients, meats, poultry and fish, or fruits (Farkas, 2006),
Due to the high value of chestnuts, it is important to develop having the additional advantage of being harmless to the environ-
conservation methodologies that allow the complete maintenance ment. However, irradiation efficacy varies significantly within dif-
of their properties. The previously applied methods include ferent fruit species, demanding continuous exposure time (doses)
fumigation (carbon disulfide, phosphine, methyl bromide), low- and geometry (dose uniformity) studies (Belchior et al., 2007;
temperature and controlled atmosphere storage, irradiation and Kim et al., 2007). In a previous study (Fernandes et al., in press),
submerging in icy water. Methyl bromide was the most widely low doses (0.27 and 0.54 kGy) of gamma irradiation were applied
used fumigant for chestnuts post-harvest disinfestation (UNEP, to chestnuts and it was found that this methodology did not affect
2006), but induces the depletion of the ozone layer and has delete- the profile and composition in important nutrients (sugars, fatty
rious effects on health, so it was banned after the Montreal Proto- acids and tocopherols). Furthermore, application of the same doses
col (Roy et al., 2008). In the European Union its use is forbidden also seemed to be advantageous for chestnuts antioxidant poten-
since March 2010 (Official Journal of the EU, 2008). Temperature tial (Antonio et al., 2011). In the present study, it is intended to
related methods may be time consuming and present low evaluate if the application of higher irradiation doses (63 kGy) still
maintain chestnuts chemical and nutritional profiles unaffected.
⇑ Corresponding author. Tel.: +351 273303219; fax: +351 273325405. Since these profiles are widely characterized (Barreira et al.,
E-mail address: iferreira@ipb.pt (I.C.F.R. Ferreira). 2009a,b; Borges et al., 2007; Borges et al., 2008; Vasconcelos

0278-6915/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fct.2011.06.062
2430 Â. Fernandes et al. / Food and Chemical Toxicology 49 (2011) 2429–2432

et al., 2007, 2010), the analyses were focused in the main compo- peaks from samples with standards. The results were recorded and processed using
CSW 1.7 software (DataApex 1.7) and expressed in relative percentage of each fatty
nents of each nutritional group: sugars – sucrose, fatty acids – pal-
acid.
mitic, oleic, linoleic and linolenic acids, tocopherols – c-tocopherol.
Furthermore, the effects of gamma irradiation on energetic contri-
2.4.3. Analysis of c-tocopherol
bution and proximate analysis of chestnuts stored at 4 °C for differ-
Tocopherols content was determined following a procedure previously de-
ent periods were evaluated, in order to understand the possible scribed by the authors (Fernandes et al., in press). The HPLC system described above
interactions among these two main factors (irradiation and storage was connected to a fluorescence detector (FP-2020; Jasco) programmed for excita-
time). tion at 290 nm and emission at 330 nm. The chromatographic separation was
achieved with a Polyamide II (250  4.6 mm) normal-phase column from YMC
Waters operating at 30 °C. The mobile phase used was a mixture of n-hexane and
2. Materials and methods ethyl acetate (70:30, v/v) at a flow rate of 1 ml/min. The compounds were identified
by chromatographic comparisons with authentic standards. Quantification was
2.1. Standards and reagents based on the fluorescence signal response, using the internal standard method.
Tocopherol contents in the samples are expressed in mg per 100 g of dry weight
Ferrous ammonium sulfate (II) hexahydrate (0.001 M), sodium chloride and sul- (dw).
furic acid (0.8 N) were purchased from Panreac S.A. (Barcelona, Spain) with purity
pa (pro-analysis). Acetonitrile 99.9%, n-hexane 95% and ethyl acetate 99.8% were
of HPLC grade from Lab-Scan (Lisbon, Portugal). The fatty acids methyl ester (FAME) 2.4.4. Statistical analysis
reference standard mixture 37 (standard 47885-U) was purchased from Sigma (St. An analysis of variance (ANOVA) with Type III sums of squares was performed
Louis, MO, USA), as also c-tocopherol and D(+)-sucrose standards. Racemic tocol, using the GLM (General Linear Model) procedure of the SPSS software, version 18.0
50 mg/ml, was purchased from Matreya (PA, USA). All other chemicals and solvents (SPSS, Inc.). The dependent variables were analyzed using two-way ANOVA, with
were of analytical grade and purchased from common sources. Water was treated the main factors ‘‘irradiation dose’’ (ID) and ‘‘storage time’’ (ST). When a (ID  ST)
in a Milli-Q water purification system (TGI Pure Water Systems, USA). was detected, the two factors were evaluated simultaneously by the estimated mar-
ginal means plots for all levels of each single factor. Alternatively, if no statistical
2.2. Samples and samples irradiation significant interaction was verified, means were compared using Tukey’s honestly
significant difference (HSD) multiple comparison test.
Chestnuts samples were obtained in an industrial unit (Agroaguiar Lda.) of Trás- In addition, a linear discriminant analysis (LDA) was used to assess the influ-
os-Montes, Northeast of Portugal. They were divided in five groups to be exposed to ence of either different storage times or irradiation doses on proximate composition
different radiation doses (0, 0.25, 0.50, 1.00 and 3.00 kGy) with 15 units per group. profiles as well as in major individual nutrients (sucrose, palmitic, oleic, linoleic and
To estimate the dose rate it was used a chemical solution sensitive to ionizing linolenic acids and c-tocopherol). A stepwise technique, using the Wilks’ k method
radiation, Fricke dosimeter, using the procedure described in a previous study with the usual probabilities of F (3.84 to enter and 2.71 to remove), was applied for
(Antonio et al., 2011). variable selection. This procedure uses a combination of forward selection and
After irradiation geometry dose rate estimation, the groups 2–5 were placed backward elimination procedures, where before selecting a new variable to be in-
into polyethylene plastic bags and irradiated with 0.25 ± 0.05, 0.50 ± 0.10, cluded, it is verified whether all variables previously selected remain significant
1.00 ± 0.20 and 3.00 ± 0.30 kGy, respectively. Group 1 was not irradiated, being (Maroco, 2003; López et al., 2008). With this approach, it is possible to identify
the control sample. Prior to analysis, all the samples were lyophilized (Ly-8-FM- the significant variables obtained for each sample. To verify which canonical dis-
ULE, Snijders, Holland). criminant functions were significant, the Wilks’ k test was applied. A leaving-one-
out cross-validation procedure was carried out to assess the model performance.
All statistical tests were performed at a 5% significance level. All the assays were
2.3. Energetic value
carried out in triplicate. The results are expressed as mean values with standard
deviation (SD).
The samples were analyzed for proximate composition (dry matter, proteins,
fat, carbohydrates and ash) using the AOAC procedures (AOAC, 1995). The crude
protein content of the samples was estimated by the macro-Kjeldahl method; the 3. Results and discussion
crude fat was determined by extracting a known weight of powdered sample with
petroleum ether, using a Soxhlet apparatus; the ash content was determined by
incineration at 600 ± 15 °C. Total carbohydrates were calculated by difference. Total Table 1 shows the proximate composition and energetic value
energy was calculated according to the following equations: data reported as mean value of each irradiation dose over three dif-
Energy ðkcalÞ ¼ 4  ðg proteins þ g carbohydratesÞ þ 9  ðg fatÞ ferent storage times, as well as mean value of five irradiation doses
within each storage time. Apart from carbohydrates, storage time
(ST)  irradiation dose (ID) interaction was a significant (P < 0.05)
source of variation for all the performed analytical assays. Hence,
2.4. Major individual nutrients
despite the least squares means are presented for the main effects,
2.4.1. Analysis of sucrose no multiple comparisons could be performed.
Free sugars were determined by high performance liquid chromatography cou- Likewise, storage time (P < 0.001) and irradiation dose
pled to a refraction index detector (HPLC-RI) as described by (Barreira et al., 2010). (P < 0.025, except for protein content) show a significant effect.
The equipment consisted of an integrated system with a pump (Knauer, Smartline
However, from the analysis of the plots of the estimated margins
system 1000), degasser system (Smartline manager 5000), auto-sampler (AS-2057
Jasco) and a RI detector (Knauer Smartline 2300). Data were analyzed using Clarity means some general conclusions can be drawn. For instance, dry
2.4 Software (DataApex). The chromatographic separation was achieved with a matter, protein and ash contents were higher for 30 days of ST,
Eurospher 100-5 NH2 column (4.6  250 mm, 5 mm, Knauer) operating at 30 °C while carbohydrates, fat and energy contents were superior in
(7971 R Grace oven). The mobile phase was acetonitrile/deionized water, 7:3 (v/
samples no submitted to storage. The different ID did not induce
v) at a flow rate of 1 ml/min. Sugar identification was made by comparing the rel-
ative retention times of sample peaks with standards. Quantification was made by
any particular tendency in the proximate composition profiles.
the internal standard method and the results are expressed in g per 100 g of dry Portuguese chestnuts chemical composition has been studied
weight (dw). by our group (Barreira et al., 2009a,b) and by other research groups
(Borges et al., 2007; Borges et al., 2008; Vasconcelos et al., 2007,
2.4.2. Analysis of palmitic, oleic, linoleic and linolenic acids 2010). Sucrose emerges as the main sugar (97% to 100%), palmitic
Fatty acids were determined by gas-liquid chromatography with flame ioniza-
(14% to 20%), oleic (23% to 31%), linoleic (42% to 52%) and linolenic
tion detection (GC-FID)/capillary column as described previously by the authors
(Fernandes et al., in press). The equipment was a GC 1000 (DANI) with a split/split- (5% to 9%) acids were the most abundant fatty acids, while
less injector, a FID and a Macherey-Nagel column (30 m  0.32 mm ID  0.25 lm c-tocopherol (88% to 100%) was the main tocopherols isoform.
df). The oven temperature program was as follows: the initial temperature of the Therefore, to evaluate the effects of irradiation dose and storage
column was 50 °C, held for 2 min, then a 30 °C/min ramp to 125 °C, 5 °C/min ramp time we focused in these major molecules.
to 160 °C, 20 °C/min ramp to 180 °C, 3 °C/min ramp to 200 °C, 20 °C/min ramp to
220 °C and held for 15 min. The carrier gas (hydrogen) flow-rate was 4.0 ml/min
Table 2 shows chestnuts major individual nutrients. The non
(0.61 bar), measured at 50 °C. Split injection (1:40) was carried out at 250 °C. Fatty stored samples revealed higher palmitic acid levels and lower
acid identification was made by comparing the relative retention times of FAME linoleic acid, linolenic acid and sucrose values. In other way, the
 Â. Fernandes et al. / Food and Chemical Toxicology 49 (2011) 2429–2432 2431

Table 1
Chestnuts energetic value according with irradiation dose (ID) and storage time (ST). In the column of carbohydrates, different letters mean significant differences.

Dry matter Fat Protein Ash Carbohydrates Energetic value

(g/100 g fw) (g/100 g dw) (g/100 g dw) (g/100 g dw) (g/100 g dw) (kcal/100 g dw)
ST 0 days 51 ± 3 2.0 ± 0.5 4±1 2.0 ± 0.2 92 ± 1 a 402 ± 3
15 days 52 ± 2 1.1 ± 0.2 6±1 2.3 ± 0.2 91 ± 1 b 396 ± 2
30 days 58 ± 4 1.1 ± 0.4 7±1 2.7 ± 0.1 90 ± 1 c 394 ± 2
P-value (n = 45) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
ID 0.00 kGy 55 ± 3 1.4 ± 0.5 6±1 2.4 ± 0.4 90 ± 1 ab 398 ± 5
0.25 kGy 52 ± 2 1.6 ± 0.4 6±2 2.4 ± 0.4 90 ± 2 b 398 ± 4
0.50 kGy 56 ± 6 1.3 ± 0.4 6±2 2.3 ± 0.3 90 ± 2 ab 397 ± 4
1.00 kGy 53 ± 5 1.3 ± 0.4 6±2 2.4 ± 0.3 91 ± 1 ab 397 ± 4
3.00 kGy 53 ± 4 1.4 ± 0.4 6±2 2.3 ± 0.3 91 ± 2 a 398 ± 2
P-value (n = 27) <0.001 0.001 0.227 0.003 0.022 0.002
ST  ID P-value <0.001 <0.001 0.044 <0.001 0.113 <0.001

Table 2
Chestnuts major individual nutrients according with irradiation dose (ID) and storage time (ST).

Sucrose C16:0 C18:1 C18:2 C18:3 c-Tocopherol

(g/100 g dw) (g/100 g dw) (g/100 g dw) (g/100 g dw) (g/100 g dw) (mg/100 g dw)
ST 0 days 21 ± 3 18 ± 3 28 ± 3 44 ± 4 7±1 0.9 ± 0.2
15 days 30 ± 8 16 ± 1 25 ± 2 48 ± 2 8±1 1.0 ± 0.1
30 days 32 ± 5 15 ± 1 25 ± 3 49 ± 2 8±1 0.8 ± 0.1
P-value (n = 45) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
ID 0.00 kGy 23 ± 3 17 ± 2 27 ± 3 46 ± 3 7±1 0.8 ± 0.1
0.25 kGy 29 ± 7 17 ± 3 26 ± 3 46 ± 5 7±1 1.0 ± 0.2
0.50 kGy 35 ± 11 17 ± 3 26 ± 2 46 ± 4 7±1 0.9 ± 0.2
1.00 kGy 26 ± 5 16 ± 2 26 ± 2 47 ± 3 8±1 0.9 ± 0.1
3.00 kGy 27 ± 6 15 ± 2 25 ± 3 49 ± 3 8±1 0.8 ± 0.1
P-value (n = 27) <0.001 <0.001 0.002 <0.001 0.002 <0.001
ST  ID P-value <0.001 0.035 <0.001 <0.001 <0.001 <0.001

samples stored for a 30 days period gave lower c-tocopherol val-

ues. Furthermore, samples irradiated with 3.00 kGy demonstrated
the highest linoleic acid contents. Generally, the acquired results
showed that the gamma radiation dose used (0.25 ± 0.05,
0.50 ± 0.10, 1.00 ± 0.20 and 3.00 ± 0.30 kGy) did not produce an
obvious effect in the assayed parameters, while storage time ex-
erted more evident influence in these values.
To verify this conclusion, the results were evaluated through a
linear discriminant analysis (LDA). All independent variables se-
lected by the stepwise procedure of the discriminant analysis were
statistically significant according to the Wilks’ k test (P < 0.05).
The LDA was performed considering different sets of the as-
sayed parameters (proximate composition, individual compounds
or both components simultaneously), in order to find which one
permitted the best classification performance. For simplicity mat-
ters, only the results obtained when all parameters were consid-
ered together are presented.
Regarding storage time, the stepwise LDA resulted in a discrim-
inant model with two significant (P < 0.001 for the Wilks’ k test)
discriminant functions. These two functions explained 100.0% of
the variance of the experimental data (the first explained 82.5%
and the second 17.5%) (Fig. 1).
The first function separates primarily 0 days and 30 days
(means of the canonical variance (MCV): 0 days = 3.290, Fig. 1. Discriminant scores scatter plot for the two defined canonical functions.
15 days = 0.016 and 30 days = 3.274), and was more powerfully
correlated with ash, protein, carbohydrates and dry matter. The
second function supported the separation of 15 days from the of the samples for the original groups and 96.3% for the cross-val-
other storage times (MCV: 0 days = 0.867, 15 days = 1.746 and idation procedure.
30 days = 0.879) and showed to be more correlated with fat, c- Regarding irradiation dose, the stepwise LDA resulted in a dis-
tocopherol and linoleic acid. The model showed a very satisfactory criminant model with three significant (P < 0.005 for the Wilks’ k
classification performance allowing to correctly classifying 97.0% test) discriminant functions. These three functions explained
2432 Â. Fernandes et al. / Food and Chemical Toxicology 49 (2011) 2429–2432

Fig. 2. Discriminant scores scatter plot for the first three canonical functions.

98.6% of the variance of the experimental data (the first explained Barreira, J.C.M., Casal, S., Ferreira, I.C.F.R., Oliveira, M.B.P.P., Pereira, J.A., 2009b.
Nutritional, fatty acid and triacylglycerol profiles of Castanea sativa Mill
60.9%, the second 21.4% and the third 16.1%) (Fig. 2).
cultivars: a compositional and chemometric approach. J. Agr. Food Chem. 57,
The first function was only able to poorly separate 0.50 kGy 2836–2842.
from the remaining doses (MCV: 0 kGy = 0.811, 0.25 kGy = Barreira, J.C.M., Pereira, J.A., Oliveira, M.B.P.P., Ferreira, I.C.F.R., 2010. Sugars profiles
0.426; 0.50 kGy = 1.421; 1.00 kGy = 0.400; 3.00 kGy = 0.635), of different chestnut (Castanea sativa Mill.) and almond (Prunus dulcis) cultivars
by HPLC-RI.. Plant Foods Hum. Nutr. 65, 38–43.
being more strongly correlated with sucrose. The second and third Belchior, A., Botelho, M.L., Vaz, P., 2007. Monte Carlo simulations and dosimetric
functions demonstrated very weak discriminant power, reflected studies of an irradiation facility. Nuc. Instrum. Method Phys. Res. A 580, 70–72.
in the classification performance of the model which allowed only Borges, O.P., Carvalho, J.S., Correia, P.R., Silva, A.P., 2007. Lipid and fatty acid of
profiles of Castanea sativa Mill. chestnuts of 17 native Portuguese cultivars. J.
to correctly classifying 59.3% of the samples for the original groups Food Compos. Anal. 20, 80–89.
and 52.6% for the cross-validation procedure. Borges, O., Gonçalves, B., Carvalho, J.L.S., Correia, P., Silva, A.P., 2008. Nutritional
The obtained results seem to indicate that the irradiation quality of chestnut (Castanea sativa Mill.) cultivars from Portugal. Food Chem.
106, 976–984.
treatment did not affect the nutritional and chemical quality of FAOSTAT, Food and Agriculture Organization of the United States, 2010. Available
chestnut fruits. In order to attend food safety issues it is now from: <http://faostat.fao.org/site/339/default.aspx> [Accessed on 5 March
necessary to assess the efficacy of this methodology in order to 2011].
Farkas, J., 2006. Irradiation for better foods. Trends Food Sci. Tech. 17, 148–152.
recommend its application as a useful alternative. Fernandes, A., Antonio, A.L., Barros, L., Barreira, J.C.M., Bento, A., Botelho, M.L.,
Ferreira, I.C.F.R., in press. Influence of low dose gamma irradiation on the profile
Conflict of Interest and composition of chestnuts (Castanea sativa Miller) in sugars fatty acids and
tocopherols. J. Agr. Food Chem.
Jermini, M., Conedera, M., Sieber, T.N., Sassella, A., Schärer, H., Jelmini, G., Höhn, E.,
The authors declare that there are no conflicts of interest. 2006. Influence of fruit treatments on perishability during cold storage of sweet
chestnuts. J. Sci. Food Agr. 86, 877–885.
Kim, J., Moreira, R., Huang, Y., Castell-Perez, M., 2007. 3-D dose distributions for
Acknowledgements optimum radiation treatment planning of complex foods. J. Food Eng. 79, 312–
321.
The authors thank to Agroaguiar Lda. for samples providing, López, A., García, P., Garrido, A., 2008. Multivariate characterization of table olives
according to their mineral nutrient composition. Food Chem. 106, 369–378.
ON.2/QREN/EU Project no. 13198/2010, for financial support. A.L. Maroco, J., 2003. Análise Estatística, com utilização do SPSS. Edições Sílabo, Lisboa,
Antonio and J.C.M. Barreira also thank to FCT – Foundation for Sci- Portugal.
ence and Technology (Portugal) for their Ph.D. (SFRH/PROTEC/ Official Journal of the European Union, 2008. Comission Decision, 753/2008, 26-09-
2008.
67398/2010) and Post-PhD (SFRH/BPD/72802/2010) Grants,
Portuguese Agricultural Statistics, 2009. Instituto Nacional de Estatística I. P.,
respectively. Lisboa, Portugal.
Roy, P., Umehara, H., Nakamura, N., Nei, D., Orikasa, T., Kitazawa, H., Okadome, H.,
Ishikawa, Y., Iwaki, K., Kobayashi, M., Shiina, T., 2008. Determination of
References
physicochemical properties of chestnuts. J. Food Eng. 87, 601–604.
UNEP, Montreal Protocol on substances that deplete the ozone layer, 2006. Report
Antonio, A.L., Fernandes, Â., Barreira, J.C.M., Bento, A., Botelho, M.L., Ferreira, I.C.F.R., of the Methyl Bromide Technical Options Committee, 205–206, 310–313.
2011. Influence of gamma irradiation in the antioxidant potential of chestnuts Vasconcelos, M.C.M.B., Bennet, R.N., Rosa, E.A.S., Cardoso, J.V.F., 2007. Primary and
(Castanea sativa Mill.) fruits and skins. Food Chem. Toxicol. doi: 10.1016/ secondary metabolite composition of kernels from three cultivars of Portuguese
j.fct.2011.02.016. chestnut (Castanea sativa Mill.) at different stages of industrial transformation..
AOAC, 1995. Official Methods of Analysis, sixteenth ed. Association of Official J. Agr. Food Chem. 55, 3508–3516.
Analytical Chemists, Arlington VA, USA. Vasconcelos, M.C.B.M., Bennett, R.N., Rosa, E.A.S., Ferreira-Cardoso, J.V., 2010.
Barreira, J.C.M., Alves, R.C., Casal, S., Ferreira, I.C.F.R., Oliveira, M.B.P.P., Pereira, J.A., Composition of European chestnut (Castanea sativa Mill.) and association with
2009a. Vitamin E profile as a reliable authenticity discrimination factor health effects: fresh and processed products.. J. Sci. Food Agr. 90, 1578–1589.
between chestnut (Castanea sativa Mill.) cultivars.. J. Agr. Food Chem. 57,
5524–5528.