Influence of the growing degree-days on chemical and technological properties of chestnut fruits (var. “Judia”)

Abstract

The influence of different edaphoclimatic conditions in “Judia” variety from seven different localities in the Trás-os-Montes region was evaluated with respect to metabolites, fatty acids, and technological properties. Significant correlations between edaphoclimatic conditions and moisture content, crude protein and especially fatty acids were investigated. Alfândega da Fé, with 2032 °D (degree-day), has the best corrected caliber and the lowest percentage of depth kernel wrinkles. However, the coldest localities (1597 °D and 1971 °D) had the highest moisture, crude protein, soluble sugars, and starch content and a clear prevalence of polyunsaturated fatty acids. The moisture loss of the chestnut fruits was also different among localities, which could lead to a different economic value. This study is important because it supplies the knowledge that trees from the same variety can produce different quality fruits depending on their origin of growth and the influence of different edaphoclimatic conditions.

La influencia de diferentes condiciones edafoclimáticas en la variedad “Judia” de siete diferentes localidades de la región de Trás-os-Montes fue evaluada con respecto a los metabolitos, los ácidos grasos y propiedades tecnológicas. Se encontraron correlaciones significativas entre las condiciones edafoclimáticas y contenido de humedad, proteína cruda y, sobre todo, en los ácidos grasos. En Alfândega da Fé, con 2032 °D, se produjeron las castañas con el mejor calibre corregido y el porcentaje más bajo de las arrugas de profundidad. Sin embargo, las localidades más frías (1597 °D y 1971 °D) presentaron el mayor contenido de humedad, proteína cruda, azúcares solubles y contenido de almidón, y un claro predominio de ácidos grasos poliinsaturados. La pérdida de humedad de los frutos de castaño (con cáscara) también fue diferente entre localidades, lo que puede repercutir en un menor valor económico de la producción. Este estudio es importante porque nos da a conocer que árboles de la misma variedad presentan frutos de diferente calidad en función de su origen y de la influencia de diferentes condiciones edafoclimáticas.

Keywords:

• chemical composition
• fatty acids
• market characteristics
• diffusivity
• multivariate analysis

Palabras clave:

• composición química
• ácidos grasos
• características del mercado
• difusividad
• análisis multivariado

Introduction

In recent years, chestnuts have become a more important part in human health because of their nutritional qualities and potential beneficial health effects. These have low protein and fat contents (99% fat free and no cholesterol), and are mostly composed of unsaturated long chain fatty acids (Borges, Soeiro Carvalho, Reis Correia, & Paula Silva, 2007). Almost half of the fresh weight of chestnut kernel is made up of water, with carbohydrates being the major component of their dry matter (almost 50%). They also contain starch and sucrose (Bernárdez, Miguélez, & Queijeiro, 2004), thus making them a healthy food (Mujic et al., 2009). Chestnuts also present significant levels of Mg, Mn, K, Cu, P, and Fe, and also vitamin C, vitamin B6, thiamine, folate, and riboflavin (De Vasconcelos et al., 2007; Mujic et al., 2009; Pereira-Lorenzo et al., 2006). Since consumers prefer the sweetest fruits, sucrose is one of the most important parameters for the assessment of the commercial quality of the fruit (Bernárdez et al., 2004). Trás-os-Montes is the main chestnut-producing region of Portugal, covering almost 85% of the total national production. This could be attributed to the excellent climatic, edaphic, and ecological conditions (Borges et al., 2007) and the predominant variety for its better fruit size (50–80 fruits/kg) (Dinis, Ferreira-Cardoso, Peixoto, Costa, & Gomes-Laranjo, in press). This variety is included in two protected designations of origin (PDO): “Castanha da Terra Fria” and “Castanha da Padrela” (Borges et al., 2007). “Judia” also predominates in new orchards, mainly due to the excellent characteristics for transformation and fresh consumption. Soil type and climatic conditions can influence the quality of the fruits (Ferreira-Cardoso & De Vasconcelos, 2009). Low water soil availability and high temperatures are regarded as the destabilizing factor for normal chestnut growth, inducing a loss of plant vigor (Gomes-Laranjo et al., 2004).

The aims of our present work were to study the chemical composition and technological properties of the chestnut fruit (var. “Judia”) grown in seven different ecotypes to compare the influence of the different climatic conditions.

Material and methods

Plant material and growth conditions

This study was carried out during 2008 on 25 chestnut trees located in seven chestnut-producing areas of Trás-os-Montes region as previously reported (Table S1; Dinis et al., in press).

The study was performed between May and October (2008). Edaphoclimatic conditions are presented in Tables S1 and S2, using methodologies already described by Dinis et al., in press.

Chemicals

All chemicals and reagents were of analytical grade (Sigma/Aldrich, Merck, and Pronalab). The water used was ultrapure and all solvents used in the determination of free sugars and fatty acids were of high-performance liquid chromatography (HPLC) grade.

Proximate composition analysis and free sugars

Dry matter, organic matter, and ash contents were determined as described by the official methods of analysis (AOAC, 1975). Crude protein and crude fat were also determined using the official methods of analysis (AOAC, 1990). Crude protein was calculated from nitrogen, determined by micro-Kjeldahl method with a selenium catalyst, by the use of factor N × 5.3 as recommended by McCarthy and Meredith (1988). Starch extraction and quantification, the initial conversion of starch to glucose (Rasmussen & Henry, 1990), was done in two phases, using a thermo-stable α-amylase (A3306-Sigma) to break it into dextrins, oligosaccharides, and amyloglucosidase (11976223-Roche) to ensure a more effective quantitative conversion to glucose (Salomonsson, Theander, & Westerlund, 1984). The colorimetric determination of glucose was done at 505 nm using the single-solution reagent method previously reported, which involves the coupled enzymatic glucose oxidase/peroxidase reaction in combination with the 4-amino antipyrine chromogen system (Blakeney & Matheson, 1984). The soluble sugars were determined at 625 nm by the method of anthrone (Irigoyen, Emerich, & Sánchez-Diáz, 1992). Free sugars were analyzed using a method based on Bernárdez et al. (2004) with small modifications. Samples (100 mg, in duplicate) were extracted with 5 mL of ethanol (60%) and 1 ml rhamnose as internal standard (25 mg/mL) for 1 h, using a horizontal agitator followed by centrifugation (3000 rpm, 5 min, 20°C). The supernatants were removed and analyzed by ion-chromatography (Dionex ICS 5000) using a CarboPac PA 20 column with a CarboPac PA 20 pre-column and detection of the sugars was done using pulse amperometric detection. The calibration curves for each of the studied sugars (fructose, glucose, sucrose, and maltose) and the quantification by comparing internal standard with the sugars in spiked samples were done.

Determination of total lipids and fatty acids composition

Total lipids were extracted with the method described by Folch, Lees, and Sloane-Stanley (1957). Fatty acids were derivatized by methylation with methanol/hexane (4:1) HCl 3 M and analyzed by gas chromatography (Slover & Lanza, 1979).

Mineral composition

Phosphorus (P) and K were analyzed after digestion with H₂SO₄ + H₂O₂ (Mills & Jones, 1996) and determined by molecular absorption spectroscopy in a segmented flow analyzer (P) or by flame emission spectroscopy (K). Ca, Mg, Zn, Cu, Fe, and Mn were analyzed after digestion with HNO₃ + HClO₄ and determined by atomic absorption spectroscopy (Mills & Jones, 1996).

Morphological and technological properties of fruits

Twenty husks were randomly collected from each tree and were stored unshelled at 2°C until analysis for a maximum of five days. Their calibers were obtained, including all yielded fruits, good, aborted, rotten, and worm fruits (total caliber, Tc) and exclusively with good fruits (corrected caliber, Cc). The pericarp and episperm were manually removed from each fruit and polyspermy was observed. To determine the kernel wrinkles, 10 fruits per tree were used. Each fruit was cut at the middle, and the width of each fruit and the wrinkle length were measured by Delta-T Devices-Windias, UK. Area Meter was used to determine the number and percentage of depth wrinkle fruits in each locality. Relative to equilibrium, moisture loss of fruits (with peel) and apparent diffusivity, chestnut fruits were harvested and stored in ventilated plastic bags at a temperature of 2°C. Each sample (five fruits per tree) was initially weighed (m ₀) at room temperature (approximately 23°C) and then placed in a plastic recipient. Each recipient, containing silica gel and a plastic grid to support the chestnuts, was inserted inside an oven at 30°C. During the trial, the samples were removed at regular intervals and weighed (m(t)) before being returned to the oven, allowing to obtain the wet-basis moisture loss at time t.

The total duration of the trial was 268 h. In this work, the kinetics of moisture loss was assumed to be governed by Fick's laws of diffusion, with constant (concentration independent) diffusivity. Moreover, chestnuts were modeled as rigid spheres of radius r given by , where a, b and c are the semi-axial dimensions of chestnuts (Cletus & Carson, 2008), initially at a uniform moisture concentration where its surface is maintained at constant concentration. Under these hypotheses, the amount of moisture loss of the chestnut is given (Crank, 1975),

where D is the apparent diffusivity and X_∞ is the moisture loss when the final equilibrium is reached. In the early stages of diffusion, the series of Equation (2) has a low convergence rate. A preliminary analysis showed that with 50 terms of the series, the equation gives an accurate description of the whole diffusion process. The identification of parameters (D and X _∞) was performed using the least square method, based on the approximation of Equation (2). This identification method is required because the drying time was not extended until no measurable weight loss was observed, inhibiting the direct measurement of X _∞. Furthermore, the applied identification method makes use of the full experimental dataset.

Statistical analysis

The results are presented as mean values ± standard deviation. For this purpose, Stat View 4.0 was used for calculation of means values, standard deviation and Fisher's test 95%. Principal component analysis (PCA) and Kruskal–Wallis ANOVA and median test (used for the determination of statistical differences between localities group) were performed using STATISTICA8 Program.

Results and discussion

Edaphoclimatic conditions

Climatic conditions are illustrated in Table S1. Concerning the mean values of temperature, Valpaços showed the lowest value, 14.7°C, and Alfândega da Fé was the hottest locality with 17.6°C. With regard to degree-day (°D) the values are quite different; Murça was 2077°D (hottest locality) and Valpaços 1597°D (coldest locality). Regarding precipitation, the values ranged between 251 mm and 38 mm (Murça and Valpaços, respectively). The mean temperature in 2008 was 1959°D. This value is quite lower than those from 2006 and 2007 (2551°D and 2155°D, respectively) (Dinis et al., in press).

Soil analysis is shown in Table S2 as already described (Dinis et al., in press).

Proximate composition analysis and free sugars

Table S3 shows the proximate composition analysis of the chestnuts studied. Moisture was the main component of chestnut, with an average content of 527.5 g/kg. Chestnut fruits from Murça (hottest locality) showed the lowest moisture value (483.1 g/kg), while chestnut fruits from Valpaços, the coldest locality, showed the highest moisture content (566.7 g/kg). The principal soil characteristics for the water availability (texture, organic matter, OM and pH_H2O) of these two localities (Murça and Valpaços) are similar (Table S2), so the differences observed between the moisture content of chestnuts could be related to the temperature sum (°D). The moisture content of the chestnut fruits harvested in 2008, are about 10% higher than the value found for the chestnuts harvested during 2006, for the same localities (Dinis et al., in press). This difference is due to the influence of the temperature sum observed during 2008 (1959°D) in comparison with them observed in 2006 (2551°D) and 2007 (2155°D).

The average starch content of chestnut fruits was 505.3 g/kg DM (Table S3). Notwithstanding the different edaphoclimatic conditions observed between the different localities selected, there were not significant differences in starch content (P > 0.05, Table S3) as occurred in 2006 and 2007 (Dinis et al., in press). However, the coldest localities have the high available carbohydrates content (starch + soluble sugars), which agree with moisture content. It is well known that both, starch and soluble sugar contents make the nuts sweeter and tastier and increased their commercial value (De Vasconcelos et al., 2009). Although, high contents of starch and moisture are responsible for mould development, which reduces the storage life time (Borges et al., 2007). The average total soluble sugars content in this study was 97.7 g/kg DM, which is lower when compared with the values of other authors (Pereira-Lorenzo et al., 2006), who referred an averaged content of 137.0 g/kg DM, for chestnuts from Galicia region (North Spain) located at higher latitude than Trás-os-Montes. In fact, in a previous study (Dinis et al., in press) it was shown that biochemical characteristics could be influenced by the edaphoclimatic conditions. The sucrose content of chestnut fruits was significantly different (P < 0.01) among localities. The sucrose content ranged between 51.4 g/kg DM for Valpaços chestnut and 98.1 g/kg DM for Chaves chestnut (1971 °D) (Table S3). The average content of glucose and fructose was 2.1 g/kg DM and 1.1 g/kg of DM, respectively which are in agreement with the values found in previous studies (Bernárdez et al., 2004). Both reductant sugars changed among localities, the glucose content variation being statistically significant (P < 0.05). Chestnut fruits from the hottest locality (Murça) contained the lowest moisture value and had one of the lowest content of soluble sugar. Nonetheless, no significant correlation was observed between total sugar content and the edaphoclimatic conditions. Protein content of “Judia” chestnut fruits ranged between 42.2 g/kg of DM and 69.5 g/kg DM in Alfândega da Fé and Valpaços, respectively. Nuts from Valpaços (the coldest) presented the highest protein content, and the protein content of chestnuts from Alfândega da Fé and Murça, the hottest localities, was the lowest one as obtained in 2006 and 2007 (Dinis et al., in press) (Table S3). However, the climatic conditions in 2007 seem to be the best such the high content of protein was observed in this year. Crude fat content presented significantly different among localities (P < 0.01) (Table S3), whose values ranged between 21.5 g/kg DM in Valpaços (coldest locality) and 44.7 g/kg DM in Murça (hottest locality). Variety “Judia” from Trás-os-Montes showed a high content in crude fat when compared to other values (Borges, Gonçalves, Carvalho, Correia, & Silva, 2008) and lower compared with values reported by De Vasconcelos et al., (2007) for the same variety. Once again the average value of crude fat in 2007 (26.8 g/kg of DM) is low compared with 2006 (28.7 g/kg of DM) (Dinis et al., in press) and 2008 (Table S3).

Minerals composition

In spite of the low ashes content, chestnut fruits have an important mineral content (Borges et al., 2008). The average content of ashes was 22.3 g/kg DM (Table 1), which was higher than the value (19.0 g/kg DM) reported by other authors for “Judia” (Borges et al., 2008), but similar to those obtained by several authors (De Vasconcelos et al., 2007; Ertürk et al., 2006; Pereira-Lorenzo et al., 2006). The content of the eight most important minerals in each ecotype was determined (Table 1). Concerning the macronutrients, potassium was the most abundant, ranging between 9.4 g/kg DM in Alfândega da Fé and 7.4 g/kg DM in Murça, suggesting that the climatic and soil mineral composition (Tables S1 and S2) induced any relation. The second most important mineral was phosphorus which ranged between 1.59 g/kg in Valpaços (coldest locality) and 1.06 g/kg DM in Murça. Phosphorus was the one which has shown the highest variability among different localities (CV = 13.8%). On the contrary of K, this nutrient seems to be quite influenced by soil composition, since soils from Valpaços presented the highest content of P (P₂O₅ = 0.23 g/kg) and Murça the poorest one (P₂O₅ = 0.028 g/kg). Calcium and magnesium contents were similar in each ecotype. Calcium contents ranged from 0.70 g/kg in Chaves to 0.55 g/kg DM in Alfândega da Fé, while Mg content varied between 0.66 g/kg in Murça and 0.54 g/kg DM in Bragança. Except calcium, all other macronutrient values are similar than those reported in other studies (Borges et al., 2008) for this same variety. Nevertheless, several authors (Pereira-Lorenzo et al., 2006; Peña-Méndez et al., 2008), in Spain, found chestnut fruits containing Mg in double the quantity than Ca. These differences could be explained by the poor content of Mg soils in Trás-os-Montes soils. Analyzing the results, a correlation was found between P, protein content and temperature which is in agreement with the data from Pereira-Lorenzo et al., (2006). It is quite evident that, the highest content of one metal can antagonize the absorption of the other metals. Consequently, the ratio between the mineral is more important than the concentrations of an individual mineral. Concerning this study, the mean value for Ca:P ratio was 0.52, which was higher than that reported by Borges et al., (2008). Our values ranged between 0.39 and 0.65 in Valpaços and Murça (coldest and hottest locality), respectively. The observed differences could be justified with the fact that soils from the warm localities are less washed and less acidic, consequently there is a great availability of these minerals.

Table 1. Mineral content (g/kg DM) of “Judia” variety located in seven localities.

Tabla 1. Contenido de los minerales (g kg⁻¹ MS) de la var. “Judia” en siete localidades.

Local	K	P	Ca	Mg	Cu	Fe	Mn	Zn	Ca:P
Alfândega da Fé	9.4 ± 0.96 a	1.13 ± 0.10 cd	0.55 ± 0.06 a	0.60 ± 0.05 a	0.0227 ± 0.001 c	0.022 ± 0.004 bc	0.041 ± 0.007 a	0.009 ± 0.17 d	0.48 ± 0.06
Bragança	8.7 ± 0.44 ab	1.15 ± 0.04 cd	0.69 ± 0.17 a	0.54 ± 0.07 a	0.0034 ± 0.002 bc	0.034 ± 0.022 abc	0.020 ± 0.005 bc	0.013 ± 0.22 c	0.60 ± 0.15
Chaves	8.1 ± 1.38 ab	1.28 ± 0.13 bc	0.70 ± 0.17 a	0.61 ± 0.08 a	0.0041 ± 0.001 bc	0.049 ± 0.031 ab	0.026 ± 0.009 b	0.014 ± 0.17 bc	0.55 ± 0.14
Mac. Cavaleiros	8.4 ± 1.85 ab	1.30 ± 0.26 bc	0.63 ± 0.12 a	0.58 ± 0.08 a	0.0034 ± 0.001 bc	0.061 ± 0.014 a	0.013 ± 0.004 c	0.016 ± 0.26 ab	0.48 ± 0.09
Murça	7.4 ± 0.49 b	1.06 ± 0.12 d	0.69 ± 0.18 a	0.66 ± 0.14 a	0.0040 ± 0.001 b	0.040 ± 0.030 abc	0.023 ± 0.008 bc	0.015 ± 0.14 abc	0.65 ± 0.07
Valpaços	9.3 ± 1.36 a	1.59 ± 0.13 a	0.61 ± 0.04 a	0.65 ± 0.16 a	0.0051 ± 0.002 ab	0.034 ± 0.021 abc	0.020 ± 0.004 bc	0.017 ± 0.17 a	0.39 ± 0.05
Vinhais	8.8 ± 0.76 ab	1.35 ± 0.08 b	0.61 ± 0.14 a	0.59 ± 0.03 a	0.0066 ± 0.002 a	0.021 ± 0.005 c	0.030 ± 0.010 b	0.015 ± 0.05 ab	0.45 ± 0.08
P	ns	**	ns	ns	*	ns	**	***	**
Mean	8.56	1.27	0.64	0.60	0.004	0.037	0.02	0.014	0.52
SD	0.69	0.17	0.06	0.04	0.00	0.01	0.01	0.003	0.10
CV (%)	8.1	13.8	8.7	7.2	30.8	38.9	35.5	18.3	18.8
Note: Values are average of all trees of every local, analyzed in triplicate±SD. The CV parameter indicates the coefficient of variation, SD the standard deviation, and DM the dry matter. Mediums within each column with the same letter are not significantly different according to the Fisher test, P ≤ 0.05 (n = 4). SL – significant level; ns – not significant (P > 0.05); *, ** and ***, significant at P < 0.05, P < 0.01, and P < 0.001, respectively.

Chestnut fruits also contain significant amounts of Zn (0.014 g/kg DM) which are higher than those (0.012 g/kg DM) found in other studies (Pereira-Lorenzo et al., 2006). With regard to micronutrients (Table 3), iron (Fe) is the most abundant mineral, with 0.037 g/kg DM followed by manganese, 0.020 g/kg DM; zinc, 0.014 g/kg DM; and copper, 0.004 g/kg DM. Except iron, which has the highest CV value (38.9%), significant differences were found between localities for the rest of the micronutrients. Zinc content is similar than that obtained by Borges et al., (2008) for “Judia” variety, but all other micronutrients are lower than those obtained by several authors (Borges et al., 2008; Portela et al., 1999).

Table 2. Total lipids and fatty acids (g/kg DM) and ΣSFA and ΣUSFA (ΣMUFA and ΣPUFA) composition (%) of the seven locals under study.

Tabla 2. Total de lípidos y ácidos grasos (g/kg MS) y composición del ΣSFA y ΣUSFA (ΣMUFA y ΣPUFA) (%) de las siete localidades estudiadas.

Local	Total Lipids	Total fatty acids	C14:0	C15:0	C16:0	C18:0	C20:0	C16:1	C18:1	C20:1	C18:2	C18:3	ΣSFA	ΣMUFA	ΣPUFA	ΣUSFA
Alfândega da Fé	33.9 ± 0.21 b	95.8 ± 8.67 a	4.1 ± 1.53 ab	0.14 ± 0.12 a	33.8 ± 4.8 abc	1.14 ± 0.82 b	2.33 ± 1.09 b	0.69 ± 0.35 b	38.8 ± 2.5 a	2.72 ± 2.14 a	14.0 ± 1.5 b	0.93 ± 1.15 a	41.5 ± 7.0 a	42.2 ± 3.6 a	14.9 ± 2.6 b	57.1 ± 3.5 ab
Bragança	35.6 ± 1.26 b	97.0 ± 0.47 a	4.4 ± 1.35 ab	0.21 ± 0.07 a	34.9 ± 1.3 bc	1.26 ± 0.74 ab	2.81 ± 1.02 ab	0.60 ± 0.19 b	40.0 ± 6.5 ab	1.39 ± 0.81 ab	12.4 ± 7.1 b	1.15 ± 1.26 a	43.5 ± 3.8 a	42.0 ± 6.7 a	13.5 ± 3.0 b	55.5 ± 4.1 ab
Chaves	36.4 ± 0.73 b	98.4 ± 0.15 a	3.8 ± 2.24 ab	1.94 ± 2.98 a	31.1 ± 4.4 c	0.90 ± 0.86 b	1.82 ± 0.96 b	0.38 ± 0.23 b	35.7 ± 5.5 ab	1.24 ± 1.15 ab	20.1 ± 9.8 ab	2.31 ± 1.39 a	38.6 ± 5.1 b	37.3 ± 6.4 ab	22.4 ± 9.2 ab	59.8 ± 5.2 a
Mac. Cavaleiros	28.1 ± 1.05 bc	97.7 ± 0.36 a	3.4 ± 0.76 b	0.38 ± 0.07 a	34.3 ± 1.6 bc	3.12 ± 0.52 a	1.66 ± 0.46 bc	1.66 ± 0.12 a	32.0 ± 5.1 ab	1.07 ± 0.83 ab	18.1 ± 7.5 ab	1.90 ± 0.07 a	43.0 ± 2.9 b	34.8 ± 6.0 ab	20.0 ± 3.4 ab	54.7 ± 3.3 ab
Murça	50.8 ± 1.10 a	98.2 ± 0.91 a	6.2 ± 0.067 a	0.72 ± 0.03 a	39.2 ± 0.3 ab	1.71 ± 2.31 ab	4.34 ± 1.50 a	1.88 ± 0.02 a	37.9 ± 1.4 b	0.67 ± 0.03 ab	10.9 ± 1.3 ab	1.75 ± 0.01 a	52.2 ± 1.8 a	40.4 ± 1.3 b	12.6 ± 1.4 ab	53.0 ± 2.7 b
Valpaços	22.6 ± 0.84 c	97.0 ± 1.21 a	2.6 ± 0.75 b	0.31 ± 0.09 a	35.1 ± 5.8 abc	2.01 ± 0.40 a	1.18 ± 0.17 c	2.01 ± 0.40 a	25.6 ± 4.1 b	0.86 ± 0.31 b	25.4 ± 5.2 a	1.93 ± 0.49 a	41.2 ± 8.0 b	28.5 ± 3.5 b	27.3 ± 5.6 a	55.8 ± 9.0 ab
Vinhais	31.9 ± 1.17 bc	95.7 ± 0.36 b	4.2 ± 0.95 ab	0.51 ± 0.06 a	40.0 ± 1.6 a	0.71 ± 0.11 b	2.94 ± 1.26 ab	1.95 ± 0.28 a	30.2 ± 6.3 ab	0.85 ± 0.32 b	12.2 ± 8.5 ab	2.10 ± 0.65 a	48.3 ± 3.6 a	33.0 ± 6.8 ab	14.3 ± 8.4 ab	47.4 ± 3.8 b
P	**	**	ns	ns	ns	ns	*	***	**	ns	**	ns	*	*	*	ns
Mean	34.19	97.1	4.10	0.60	35.5	1.55	2.44	1.31	34.3	1.26	16.1	1.73	44.0	36.9	17.8	54.8
SD	8.75	1.07	1.09	0.62	3.12	0.83	1.05	0.72	5.23	0.69	5.30	0.50	4.65	5.11	5.50	3.87
CV (%)	25.6	1.1	26.6	103.0	8.8	67.6	33.8	55.0	15.2	55.0	32.9	29.1	10.6	13.8	30.9	7.1
Note: Values are average of all trees of every local±SD. The CV parameter indicates the coefficient of variation and SD de standard deviation. Mediums within each column with the same letter are not significantly different according to the Fisher test, P ≤ 0.05 (n = 4). SL – significant level; ns – not significant (P >  0.05); *, **, and ***, significant at P < 0.05, P < 0.01, and P < 0.001, respectively. SFA = saturated fatty acids; USFA = unsaturated fatty acids; MUFA = monounsaturated fatty acids; PUFA = polyunsaturated fatty acids.

Table 3. Morphological fruits parameters of “Judia” variety evaluated in seven localities from Trás-os-Montes region.

Tabla 3. Parámetros morfológicas de los frutos de la variedad “Judia” evaluados en siete localidades de la región de Trás-os-Montes.

	Calibre (fruits/kg)			Polyspermy	Depth wrinkles		Moisture loss
Local	Corrected (Cc)	Total (Tc)	Tc/Cc	(fruits/kg)	(%)	N° wrinkles	X∞(%)
Alfândega da Fé	60.7 ± 5.2 d	78.7 ± 10.4 d	1.3 ± 0.07	3.1 ± 3.3 a	15.2 ± 4.9 b	11.0 ± 1.4 b	45.2 ± 1.9 cd
Bragança	74.8 ± 10.0 bc	115.1 ± 32.4 c	1.5 ± 0.26	1.3 ± 2.4 b	21.0 ± 3.7 a	12.0 ± 4.2 b	49.8 ± 1.9 b
Chaves	63.2 ± 4.8 ab	103.0 ± 7.2 c	1.6 ± 0.17	1.8 ± 2.1 a	1.8 ± 2.1 ab	16.0 ± 4.9 a	47.2 ± 2.0 c
Mac.Cavaleiros	107.1 ± 10.7 a	291.1 ± 19.1 a	2.7 ± 0.10	0.0 ± 0.0 b	18.0 ± 2.6 ab	12.3 ± 0.6 b	52.8 ± 1.8 b
Murça	89.6 ± 2.6 ab	169.4 ± 58.4 b	1.9 ± 0.71	1.0 ± 1.6 a	18.7 ± 0.6 ab	13.5 ± 0.7 ab	43.8 ± 2.2 d
Valpaços	66.0 ± 3.5 cd	104.1 ± 8.6 cd	1.6 ± 0.08	4.7 ± 4.6 a	20.9 ± 3.1 a	16.5 ± 1.7 a	53.6 ± 1.7 a
Vinhais	67.5 ± 3.8 cd	108.5 ± 3.5 cd	1.6 ± 0.22	2.8 ± 4.8 a	19.5 ± 22 ab	17.8 ± 4.5 a	50.5 ± 0.6 b
P	***	***		***	ns	ns	***
Mean	75.6	138.6	1.75	2.09	18.7	14.2	49.0
SD	16.9	72.7	0.46	1.55	2.02	2.59	3.73
CV (%)	22.4	52.5	26.3	73.9	10.8	18.3	7.6
Note: Values are average of all trees of every local±SD. Within each column, means with the same letter are not significantly different, according to the Fisher test (P > 0.05); ns – not significant (P > 0.05); ** and ***, significant at P < 0.01 and P < 0.001, respectively.

Total lipids and fatty acid composition

The lipid composition of chestnut fruits is described in Table 2. The average of total lipids content obtained for all localities under study was 34.19 g/kg DM (Table 2), which is higher than those obtained by several authors for C. Mollissima (Senter et al., 1994) and Spanish varieties (30 g/kg DM) (Pereira-Lorenzo et al., 2006). Total lipids varied between 22.6 and 50.8 g/kg DM, for the coldest and hottest locality, respectively. The lipid fraction was composed of 44.0% of saturated fatty acids (SFA), 54.8% unsaturated fatty acids (USFA), 36.9% monounsaturated fatty acids (MUFA) and 17.8% of polyunsaturated fatty acids (PUFA) (Table 2). The content of SFA was 17.8% higher than those reported by other studies done during the cold years of 2001 and 2002, for “Judia” (Borges et al., 2007). It is recognized that acclimation to high temperatures is normally associated with a greater degree of saturation of fatty acids (Hardwood, 1998). Nevertheless the results are in agreement with our work compared with the results of the selected localities; Murça (hottest locality) has significantly higher amounts of SFA, whereas Chaves and Valpaços (coldest localities) had the lowest SFA value. Palmitic acid (C16:0) is clearly the most important SFA in chestnut nut as observed in other works (Borges et al., 2007). The total PUFA content ranged between 12.6% (Murça) and 27.3% (Valpaços) with linoleic acid (C18:2) being the most important. The content of USFA was 36.1% lower than those reported by other studies (Borges et al., 2007; Senter et al., 1994). Concerning total MUFA content, the values ranged between 42.2% and 28.5%, showing significant differences between the different studied sites, which is in agreement with the results reported in previous studies (Borges et al., 2007). These results emphasize a clear correlation between PUFA and °D. In fact, fruits from coldest localities (Valpaços and Chaves) show higher PUFA content. This increase of PUFA as a response to temperature diminishment is a well-known phenomenon where cell membrane fluidity could be adjusted by changing the PUFA content.

Morphological and technological properties of fruits

Due to the size of its fruits, “Judia” is very well appreciated in the market, both as fresh and also in confectionery industries. Concerning calibers, significant values were obtained among the localities (P < 0.001) (Table 3). Fruits produced in Alfândega da Fé were the biggest, having a total caliber of 78.7 fruits/kg, while in Macedo de Cavaleiros (2050 °D) a total caliber of 291.1 fruits/kg was measured. Subtracting the aborted and infested fruits (corrected caliber), fruit caliber ranged between 60.7 fruits/kg at Alfândega da Fé, and 107.1 fruits/kg at Macedo de Cavaleiros. During 2008 (May–October: 1959°D), the average value for the corrected caliber was 75.6 fruits/kg, which indicates a difference of 63 fruits/kg in relation to the total caliber. In the Portuguese context, “Judia” produces the biggest fruits. According to Ferreira-Cardoso (2007) the oldest Portuguese variety Longal only produces fruits with a mean caliber of 115.8 fruits/kg. In Spain, Pereira-Lorenzo et al., (2006) found, among 47 varieties, a range between 44 fruits/kg for Tomasa (Andalucía) and 155 fruits/kg for Del Haya (Canary Islands). The smallest ratio of total/corrected caliber was obtained in Alfândega da Fé, which could be justified both, with the local climatic conditions and presence of other chestnut genotypes able to pollinate “Judia”. The coefficient of variation was considerably higher for total caliber (CV = 52.5%) compared to that obtained for corrected caliber (CV = 22.4%). In any case, these CV values are higher than 20%, which indicates a bigger variability in this morphologic characteristic within localities.

Another important aspect which valorises the production is the low level of polyspermy percentage (Ferreira-Cardoso and De Vasconcelos, 2009). The highest polyspermy value (4.7 fruits/kg) was observed in the coldest locality, while in Macedo de Cavaleiros we did not obtain any polyspermy in their fruits (Table 3). The overall value was 2.09 fruits/kg, which is lower than 4.65% obtained in a previous study for the same variety (Pimentel-Pereira et al., 2007). These low values allow classifying this variety “marron” characteristic since it presents polyspermy degree lower than 12% (Bergougnoux, 1978). Several studies have demonstrated the frequent occurrence of kernel wrinkles (grooves) on chestnut fruits (Ferreira-Cardoso et al., 1993; Pereira-Lorenzo et al., 1996). Results from different “Judia” ecotypes (Table 3), are in agreement with previously observed ecotypes (Bolvanský, Brindza, and Horcin, 2005) for wrinkles kernels. Our data show that the highest percentage of depth wrinkles were obtained in Bragança and Valpaços; however, the highest number of wrinkles was found in Vinhais and Valpaços. “Judia” variety presented elevated levels of depth wrinkles (grooves) compared with other Portuguese cultivars such as Lada and Benfeita (Ferreira-Cardoso et al., 1993). The presence of wrinkles can increase the fragmentation of the fruit during processing, which is one of the criteria used for fruit selection. Several studies (Ferreira-Cardoso et al., 1993; Bolvanský et al., 2005) have shown a correlation between fruit size and wrinkles and concluding that, the biggest fruits have the highest number and depth of kernel wrinkles. Nevertheless, our results did not agree with this conclusion, since the biggest fruits, 60.7 fruits kg⁻¹ (Alfândega da Fé), presented the lowest number and percentage of depth wrinkles, 11 and 15%, respectively.

It was evident that the significant differences of moisture loss occurred according to their origin and climatic conditions (Table S3 and Table 3). Valpaços had the major percentage of moisture loss (53.6%), while Murça (hottest locality) had the minor percentage of moisture loss (43.8%) (Table 3). Alfândega da Fé presents significant differences in the apparent diffusivity of water (A – 4.4 × 10⁻¹¹ m²/s, B – 5.0 × 10⁻¹¹ m²/s, C – 6.0 × 10⁻¹¹ m²/s, Ma – 5.3 × 10⁻¹¹ m²/s, M – 5.5 × 10⁻¹¹ m²/s, V – 4.8 × 10⁻¹¹ m²/s and Vi – 4.8 × 10⁻¹¹ m²/s) among the selected localities. Higher moisture in fruits (with peel) is essential for storage to the fresh market, therefore a problem during storage and delivery. Despite the dry appearance, chestnuts are perishable and therefore present a limited shelf-life (Sacchetti et al., 2005). The mean apparent diffusivity obtained in the study was 5.2 × 10⁻¹¹ m²/s (with a CV = 9.9%), which is similar to the apparent diffusivity (5.1 × 10⁻¹¹ m²/s) (Cletus and Carson, 2008) in New Zealand. The loss of moisture can represent a significant loss of weight and consequently a lower economic performance.

Classification by multivariate analysis

In order to extract the most important sources of variability and variable correlations in our dataset, a PCA was performed to all variables and clusters under study. Figure 1A shows the projection of the variables in the first two PCs (42.5% of the total variance). The variables with a high positive correlation with the PC1 were ratio total/corrected caliber, polyspermy, MUFA, crude fat and degree-day, while the percentage of loss of water, moisture content, Cu, pH soil, PUFA and total caliber presented a high negative correlation with PC1. Moreover, the corrected caliber, polyspermy and soluble sugars present the largest contribution in the (+)PC2 and corrected caliber and °D in (−)PC2. An association between soil pH, total and corrected caliber was found. As can be observed (Figure 1B), five distinct groups of trees can be observed based on the projection of the cases: V, A, M and Ma in different groups and B, C and Vi are in the same group. In fact, Valpaços is the coldest locality and is distant from Murça, the hottest one. Alfândega da Fé, the second hottest locality, is close to Murça but in the other group, may be due to different edaphic conditions. Based on PCA analysis, the correlation between high moisture content and high percentage of loss water is clear. This loss is also correlated with the number and depth of wrinkles. In fact, chestnut fruits from Valpaços have the highest moisture content, high number of wrinkles and, have also the highest percentage of water loss (Figure 1A and B). Nevertheless, this ecotype presented the highest percentage of PUFA. On the other hand, fruits from Murça (Figure 1B) had the highest percentage of SFA and crude fat (Figure 1A). The PCA analysis also showed an association between the protein and phosphorus variables. Indeed, a correlation was observed between the fruits moisture and protein content and degree-day, supporting the idea that the °D could constitute the most important factor influencing this content of chestnuts. This analysis corroborates the idea that the hottest localities had more crude fat content and SFA. Also, we found a positive correlation between °D with SFA and negative correlation with total USFA, which is in concordance with other studies (Borges et al., 2007; Sánchez-Machado et al., 2004).

Figure 1. Principal component loadings plot (A) and scores plot (B) for the two first principal components for all parameters under study.

Figura 1. Diagrama de cargas de componentes principales (A) y gráfica de resultados (B) de los dos componentes principales para todos los parámetros estudiados.

Conclusion

Chestnut fruits can be a good alternative to conventional starch resources, such as cereals and tubers, making them ideal in many food and non-food applications (Correia and Beirão-da-Costa, 2009).

This study was carried out within the same variety; however, great differences were found between the selected ecotypes (localities with different edaphoclimatic conditions). It is important to refer that degree-day values seem to affect the chemical composition of fatty acids and technological properties of fruits of the “Judia” variety. Fruits from the coldest localities show high crude protein, soluble sugars, starch, the highest percentage of USFA (especially linoleic acid) and less percentage of SFA. It is necessary to emphasize the highest amount of PFA present in chestnut fruits grown in the coldest localities. However, with regard to best fruit size, (caliber) the optimal values range between 1970 and 2032°D (May−October). Based on the results of the PCA analysis we can conclude that the variation of the fruit characteristics seem to be essentially dependent on the climatic conditions since in this work we proved that all trees belong to the same variety.

Supplementary material

The supplementary material for this paper is available online at http://dx.doi.org 10.1080/19476337.2011.631713

Supplementary Table S1. Climatic, geographic (A) and edaphic (B, 0 – 50 cm) characterization of different localities under study.

Tabla adicional S1. Caracterización climática, geográfica (A) y edáficas (B, 0 – 50 cm) de las diferentes localidades en estudio.

		Coordinates		Altitude	Average Temp. (°C)	Degree-days (°D)	Precipitation (mm)
Local	No. of trees	“N°”	“W°”	(m)	(May-Oct)	(May-Oct)	(May-Oct)
Alfândega da Fé	4	41.42	6.95	759	17.62	2032	215
Bragança	4	41.68	6.72	848	16.93	2010	167
Chaves	4	41.74	7.24	709	17.43	1971	208
Mac. Cavaleiros	3	41.60	6.95	722	17.15	2013	232
Murça	2	41.52	7.41	711	17.30	2077	251
Valpaços	4	41.60	7.43	860	14.70	1597	227
Vinhais	4	41.83	7.08	816	16.93	2010	167
Mean	3.57	41.63	7.11	775	16.87	1959	210
SD	0.79	0.14	0.26	65.78	0.99	162.73	31.93

Supplementary Table S2. Proximate composition and free sugars (g/kg DM) of different locals under study.
Tabla adicional S2. Composición aproximada y azú cares libres (g/kg MS) de los diferentes locales en estudio.

Local	Texture	pHH2O	OM	P2O5	Ca	Mg	K	Na
			(%)	(g Kg^−1)	(cmol Kg^−1)	(cmol Kg^−1)	(cmol Kg^−1)	(cmol Kg^−1)
Alfândega da Fé	loam	4.68	0.64	0.053	1.34	0.32	0.34	0.08
Bragança	loam	5.10	4.19	0.110	0.99	0.26	0.32	0.08
Chaves	sandy loam	5.23	1.04	0.064	1.01	0.23	0.39	0.10
Mac. Cavaleiros	loam	5.58	2.93	0.103	2.01	0.38	0.26	0.09
Murça	loam	5.10	2.43	0.028	0.48	0.23	0.30	0.08
Valpaços	loam	5.02	2.43	0.230	0.71	0.19	0.33	0.09
Vinhais	loam	5.24	4.56	0.044	1.34	0.35	0.43	0.52
Mean		5.14	2.60	0.090	1.13	0.28	0.34	0.15
SD		0.27	1.46	0.069	0.05	0.07	0.05	0.17
CV(%)		5.3	56.1	76.2	44.3	25.1	15.9	110.0
OM: organic matter

Supplementary Table S3. Proximate composition and free sugars (g/kg DM) of different locals under study.

Tabla adicional S3. Composición aproximada y azúcares libres (g/kg MS) de los diferentes locales en estudio.

Local	Moisture	Crude protein	Crude fat	Ash	Starch	Soluble sugars	^(1)Available carbohydrates	Glucose	Fructose	Sucrose
Alfândega da Fé	499.1 ± 3.9 c	42.2 ± 3.2 c	27.0 ± 7.7 bc	24.0 ± 1.3 a	503.7 ± 29.3 a	113.6 ± 11.2 ab	61.7 ± 3.5 a	1.7 ± 0.5 b	0.8 ± 0.3 bc	81.8 ± 6.7 b
Bragança	529.1 ± 10.9 b	44.9 ± 7.5 c	30.8 ± 4.5 b	22.2 ± 1.7 b	507.8 ± 31.9 a	90.6 ± 18.8 bc	59.8 ± 2.4 ab	2.8 ± 0.6 a	1.0 ± 0.4 b	72.2 ± 9.5 c
Chaves	537.9 ± 14.5 b	49.1 ± 8.4 bc	32.1 ± 3.0 b	21.6 ± 2.5 b	498.5 ± 12.7 a	110.7 ± 15.5 ab	60.9 ± 1.1 a	2.6 ± 1.4 ab	1.3 ± 1.1 ab	98.1 ± 16.3 a
Mac. Cavaleiros	536.2 ± 16.0 b	51.3 ± 10.5 b	23.3 ± 3.9 c	21.6 ± 0.6 b	492.7 ± 29.6 a	71.5 ± 16.2 c	59.1 ± 4.0 ab	3.1 ± 0.9 a	1.9 ± 0.8 a	72.7 ± 5.4 c
Murca	483.1 ± 8.6 d	43.1 ± 0.3 bc	44.7 ± 0.1 a	20.7 ± 2.8 c	519.8 ± 22.4 a	89.8 ± 25.5 bc	58.2 ± 0.3 b	1.1 ± 0.8 c	0.5 ± 0.1 c	69.9 ± 4.8 c
Valpaços	566.7 ± 11.4 a	69.5 ± 6.4 a	21.5 ± 1.5 c	23.5 ± 2.3 ab	512.7 ± 17.4 a	120.5 ± 31.8a	63.3 ± 4.9 a	0.6 ± 0.5 c	1.0 ± 0.5 b	51.4 ± 10.1 d
Vinhais	540.7 ± 16.7 b	52.7 ± 4.3 b	27.3 ± 4.5 bc	22.4 ± 1.6 b	501.8 ± 34.8 a	87.4 ± 6.2 bc	58.9 ± 3.5 b	2.6 ± 1.6 ab	1.5 ± 0.6 a	73.5 ± 6.9 c
P	***	**	**	ns	ns	*	*	*	*	**
Mean	527.5	50.0	29.5	22.3	505.3	97.7	60.3	2.1	1.1	74.2
SD	27.9	9.7	7.7	1.1	9.1	17.5	1.8	0.9	0.5	14.0
CV(%)	5.3	19.4	26.0	5.1	1.8	18.0	3.0	45.8	40.7	18.9
Note: Values are average of all trees of every local, analyzed in triplicate±SD. The CV parameter indicates the coefficient of variation, SD the standard deviation, and DM the dry matter. Mediums within each column with the same letter are not significantly different according to the Fisher test, P ≤ 0.05 (n = 4). SL – significant level; ns – not significant (P > 0.05); *, ** and ***, significant at P < 0.05, P < 0.01, and P < 0.001, respectively. ^(1)starch + soluble sugars.

Acknowledgments

The authors are grateful to the FCT (Foundation for Science and Technology, Portugal) for financial support of the scholarship of Lia-Tânia Rosa Dinis (SFRH/BD/29220/2006). They also thank the Assistant-technician Maria Natália Campos Teixeira for the support given to the laboratorial work.