Chilean journal of agricultural research
versión On-line ISSN 0718-5839
Chilean J. Agric. Res. v.70 n.3 Chillán sep. 2010
Chilean Journal of Agricultural Research 70(3):345-356 (July-September 2010)
Adaptation and Genotype x Environment Interaction of Flaxseed (Linum usitatissimum L.) Genotypes in South Central Chile
Adaptación e Interacción Genotipo x Ambiente en Lino (Linum usitatissimum L.) en la Zona Centro Sur de Chile
Marisol Berti1*, Susana Fischer1, Rosemarie Wilckens1, Felicitas Hevia2, and Burton Johnson3
1Universidad de Concepción, Facultad de Agronomía, Av. Vicente Méndez 595, Chillán, Chile. *Corresponding author (firstname.lastname@example.org).
2Universidad de Concepción, Facultad de Ingeniería Agrícola, Av. Vicente Méndez 595, Chillán, Chile.
3North Dakota State University, Department of Plant Sciences, Fargo, ND 58108, USA.
Flaxseed (Linum usitatissimum L.) is imported into Chile mostly for bread making and feed. Identification of genotypes best adapted for seed production in South Central Chile would facilitate producer's decision. The objective of this study was to determine the adaptation and genotype x environment interaction of 16 flaxseed genotypes (including 10 from North American and six from Argentine sources) grown at 11 environments (defined as location-year) in Chile from 2003 to 2007. Genotype seed yield was above 5700 kg ha-1 for some environments indicating a high yield potential. According to the AMMI (Additive Main Effects and Multiplicative Interaction) and SREG (Sites regression) models the 11 environments were classified into four groups by the AMMI and three groups by the SREG models. Genotypes were classified into five groups by the SREG model with four of the groups as single genotypes. Overall mean seed yield was similar for all genotypes; however the genotype Nekoma was the most stable and higher yielding genotype across environments. The environment with the highest yield potential was Chillán 2003-2004, but this location had low yield stability across years. The environments with greatest seed yield potential, Chillán 2003-2004 and Los Ángeles 2004-2005, had irrigation during flowering and seed filling. Seed oil content fluctuated between 420 and 530 g kg-1. The climatic differences among environments did not influence oil composition as expected from previous research. Flaxseed appears adapted to South Central Chile with differences observed among genotypes for biomass and seed yield, harvest index, test weight, oil content, and composition.
Key words: adaptation, genotype, seed yield, oil content, linolenic acid, Linum usitatissimum.
La semilla de lino (Linum usitatissimum L.) se importa a Chile principalmente para panaderías y alimento animal. La identificación de genotipos altamente productivos en la zona Centro Sur de Chile facilitaría la decisión de los productores. El objetivo de este estudio fue determinar la adaptación y la interacción genotipo x ambiente de 16 genotipos de lino (incluyendo 10 procedentes de Norte América y 6 de Argentina) los que se sembraron en 11 ambientes (localidad-año) en Chile entre el 2003 y el 2007. El rendimiento de semillas observado fue de más de 5700 kg ha-1 en algunos ambientes lo que indicaría un alto potencial. De acuerdo a los modelos AMMI y SREG los 11 ambientes fueron clasificados en cuatro grupos por el modelo AMMI y en tres grupos por el modelo SREG. Los genotipos fueron clasificados en cinco grupos por el modelo SREG, donde cuatro de los grupos incluían un sólo genotipo. El promedio general de rendimiento de semillas fue similar para todos los genotipos, sin embargo, el genotipo Nekoma fue considerado el más estable y de mayor rendimiento de semillas en todos los ambientes evaluados. El ambiente con mayor rendimiento fue Chillán 2003-2004, pero esta localidad tuvo baja estabilidad en el tiempo. Los ambientes con más alto rendimiento fueron Chillán 2003-2004 y Los Ángeles 2004-2005, hay que considerar que ambos ambientes fueron regados durante floración y llenado de granos. El contenido de aceite fluctuó entre 420 y 530 g kg-1. Las diferencias climáticas entre ambientes no influenciaron la composición del aceite como se esperaba. El lino se adapta a la zona centro sur de Chile observándose diferencias entre genotipos para rendimiento de biomasa y semillas, índice de cosecha, peso del hectolitro, contenido y composición del aceite.
Palabras clave: Adaptación, genotipo, producción de semilla, contenido de aceite, ácido linolénico, Linum usitatissimum.
Flax belongs to the Linaceae family and its origin is in Europe and Asia (Casa et al., 1999; Charlton and Ehrensing, 2001; Berglund and Zollinger, 2002). Flax was introduced to the Americas by Spaniards, and brought to Chile at the beginning of the 19th century. Flax world production area is 2 339 210 ha. World average seed yield is 943 kg ha-1, and Canada, India, China, and the USA are the main producing countries (FAO, 2009). Canada is the main flax seed exporter followed by the USA which together have 93% of the marketed volume in the World. The European Union is an important importer of flax seed (DIMEAGRO, 2007).
In Chile, there are only 50 ha of cultivated flax, mostly located in the South Central region of Chile. Chile imports 100% of the flax seed and flax oil consumed in the country. Most imported flax seed comes from Argentina, which produced 28 000 ha of flax in the 2006-2007 seasons (DIMEAGRO, 2007). Flax seed imported into Chile is used mainly for bread. Seeds have 400 to 420 g kg-1 of oil in most countries and the main fatty acids are alpha-linolenic acid (ALA), oleic acid, and linolenic acid (LA). Flax seed oil has approximately 50% of ALA and omega-3 fatty acid. The oil has a ratio of 1:3 between omega-3 and omega-6 fatty acids which is not easily found in other seed oils (Berglund and Zollinger, 2002). Flax seeds are sold for human consumption (nutraceutical and pharmaceutical industries) and also to feed poultry (chicken) and swine in Canada and the USA. Flaxseed oil is also used as industrial oil in varnishes and paints.
The human body does not synthesize ALA and LA which are required in the diet, thus they are called essential fatty acids. The ALA is required in the body to synthesize longer chain omega-3 fatty acids such as stearidonic acid, eicosapentanoic (EPA) and docosahexaenoic (DHA) acids. Flaxseed oil is highly reactive and easily becomes rancid. Only stabilized, encapsulated flax seed oil can be consumed directly by humans. The intake of flax seed and flax seed oil in humans and animals is limited by the presence of antinutritional factors (Madhusudhan et al., 1986; Oomah et al., 1996).
Flax is adapted to many environments mainly in temperate climates (Casa et al., 1999; Adugna and Labuschagne, 2003). It has been demonstrated that seed yield and seed yield components, plant height, time to reach harvest maturity, oil content, and oil composition, depend mainly on the temperature during plant development (Dybing et al., 1988; D'Antuono and Rossini, 1995; Casa et al., 1999; Adugna and Labuschagne, 2003; Cross et al., 2003). Previous research indicated the oil content of flax seed decreased 40 gkg-1 when the day temperature increased from 15 to 27 °C and the night temperature increased from 10 to 22 ºC. Other factors such as seed moisture at harvest may also influence the final seed oil content and the oil composition (Froment et al., 1999; Adugna and Labuschagne, 2003).
The genotype x environment interaction is the result of the yield response of each genotype to the variations in soil fertility, water availability, temperature, photoperiod, light intensity, crop management, and other (Crossa et al., 1991). This interaction is a confounded effect of the genotype observed mean performance and its true value. Thus, the resulting phenotype and yield expression of a cultivated genotype will vary among environments due to their diversity of growth resources. This allows us to select the environment to which the genotype is adapted due to the differential adaptation in comparison with other genotypes evaluated simultaneously in the same environments (Romagosa and Fox, 1993; Diepenbrock et al., 1995; Adugna and Labuschagne, 2003).
In order to study the nature and magnitude of the genotype x environment (G x E) interaction and the adaptation of a genotype to a specific environment linear models (only additive effects) and bilineal models (additive and multiplicative effects) can be used. The additive main effect and multiplicative interaction (AMMI) and the sites regression (SREG) models have been used to analyze the G x E interaction (Zobel et al., 1988; Crossa et al., 1991; Crossa and Cornelius, 1997; Vargas and Crossa, 2000; Yan et al., 2000; Ma et al., 2004). The results of both models are displayed on biplots (Kempton, 1984). The AMMI results are displayed on a G x E biplot, and this enables determination of the main effect of the genotype, the environment, and the most meaningful G x E interactions. The SREG model is displayed as a GGE biplot (genotype, genotype x environment) based on the genotype main effect and the G x E interaction. This last model can be used for genotype evaluation, environment evaluation, and mega-environment analysis (Ma et al., 2004).
The objective of this study was to determine the adaptation and genotype x environment interaction of 16 flaxseed genotypes (including 10 North American and six Argentine sources) grown at 11 environments (defined as location-year) in south central Chile from 2003 to 2007.
MATERIALS AND METHODS
The experiment was conducted in Chillán, Yungay, Los Ángeles, and Osorno in the 2003-2004, 2004-2005, 2005-2006 and 2006-2007 seasons. Chillán (36°35' S, 72°04' W, 140 m.a.s.l.) is located in the Ñuble Province, Bío Bío Region, Chile. Soil at Chillán is medial, thermic Humid Haploxerands (Arrayán series), plain and good drainage and an average rainfall of 1000 mm (Carrasco, 1998). The climate of this location is classified as temperate Mediterranean and belongs to the agroclimate Chillán (Novoa et al., 1989; Del Pozo and Del Canto, 1999). Yungay (37º03' S, 72º01' W, 263 m.a.s.l.) is located in the Ñuble Province, Bío Bío Region. Soil at Yungay is ashy, mesic Typic Haploxerand (Santa Bárbara series), slightly hilly and an annual rainfall from 1200 to 1500 mm the climate is classified as cold Mediterranean and the agroclimate is high foothills (Novoa et al., 1989). Los Ángeles (37º27´ S, 72º18´ W, 210 m.a.s.l.) is located in the Bío Bío Province, Bío Bío Region. The soil at Los Ángeles is ashy, mesic Typic Haploxerand (Santa Bárbara series), slightly hilly with an annual rainfall from 1000 to 1500 mm. The climate is classified as temperate Mediterranean and belongs to Chillán agroclimate (Del Pozo and Del Canto, 1999).
Osorno is located (40º22' S, 73º04' W, 72 m.a.s.l.) in Los Lagos Region, Chile. The experiments were conducted under dryland conditions and no-till. Soil at Osorno is ashy, mesic Typic Haploxerand (Osorno series), slightly hilly and rainfall from 1200 to 1500 mm. Climate is classified a cold Mediterranean (Novoa et al., 1989). The chemical and physical soil analysis and the average monthly temperature and rainfall of each environment are indicated in Table 1.
All environments had 16 genotypes except in Yungay 2003-2004, which had only 10 genotypes. Experimental units consisted of six rows, 5 m long and spaced 20 cm apart. Ten of the genotypes tested were from the USA and six of them were from Argentina. Genotypes and their characteristics are indicated in Table 2.
Seeding dates, seeding rates, fertilizer rates, and harvest dates are indicated in Table 3. All fertilizers were applied and then incorporated in furrows at 4 cm depth at seeding. The experiments were conducted under conventional tillage and with supplemental irrigation at Chillán and Los Ángeles, and dryland conditions and no-till at Yungay and Osorno. The previous crop was wheat (Triticum aestivum L.) at all environments. Fertilizer rates were calculated at each location according to soil analysis. Broadleaf and grass weeds were controlled with pre-plant incorporation of trifluralin [2,6-dinitro-N,N-dipropyl-4-(trifluoromethyl)benzenamine] at 1 L ha-1 (Treflan, Dow Agrosciences, Chile). Post-emergence broadleaf weed control was with MCPA dimethylamine salt at 0.46 L ha-1 (MCPA 750SL, ANASAC, Chile), and for grass weeds quizalofop-p-tefuryl at 0.275 L ha-1 (Pantera Plus, Crompton Corporation, Chile). Yungay and Osorno plots were hand weeded. In Chillán in 2003-2004 and Los Ángeles in 2004-2005, flax was irrigated by furrows; irrigation frequency was every 7 d replacing about 70% of pan evaporation. The two-center rows of each plot were swathed and threshed with a stationary plot combine (Bill's Welding, Pullman, Washington, USA).
Biomass samples were taken from a 0.2 m2 area within each plot where plants were cut at the stem base. Harvest index was calculated as the percentage of dry seed weight divided by the total dry aboveground biomass from the 0.2 m2 harvested area within each plot. Biomass and harvest index were determined only in Chillán, Yungay, and Osorno in 2003-2004. One thousand seeds were counted with an electronic seed counter and weighed to determine 1000 seed weight. Test weight was calculated by determining the weight of 100 mL of seed from the harvested seed yield of each plot. Seed yield was taken from the center-two rows of each experimental unit discarding 0.5 m of plants from row ends. Seed oil content was determined on a 40 mL sample of each experimental unit with a Newport Nuclear Magnetic Resonance (NMR) analyzer, (Newport 400, Oxford Institute, Oxford, England).
Seed fatty acid composition analysis was conducted at the Facultad de Ingeniería Agrícola, Universidad de Concepción. Gas chromatography of methyl esters was performed with a gas chromatographer (Varian 3900, Varian, Palo Alto, California, USA), equipped with a flame ionization detector (FID). Analysis was conducted on a CP-WAX 52 CB 30 x 0.25 mm silica gel column. The analysis was conducted as follows: column flow 1 mL min-1 programmed ramp at 120-240 ºC in three stages 120 ºC for 3 min, and increasing at 3 ºC min-1 to 210 ºC, for 55 min, and a final increase of 15 ºC min-1 to 240 ºC for 65 min. Injector and detector temperatures were set at 200 and 300 ºC, respectively (Ackman, 2002).
Experimental design and statistical analysis
The experimental design was a randomized complete block design (RCBD) with four replicates at the 11 environments tested. Statistical analyses were conducted using standard procedures for a randomized complete block design at each environment (Steel et al., 1996). The SAS system was used to process the data (SAS Institute, 2007). Each combination of location-year was defined as an environment and considered a random effect. The fall and spring seeding at Chillán 2004-2005 were considered separate environments. Genotypes were considered fixed effects. Mean separation was performed by applying F-protected least significant difference (LSD) comparisons at P ≤ 0.05 level of significance.
Analysis of Genotype x Environment interaction
The genotype by environment data analysis was performed using the AMMI and SREG models. Two types of biplots, AMMI biplot (Zobel et al., 1988) and GGE biplot (Yan et al., 2000; 2007) were used to visualize the genotype by environment interaction. For this analysis the 10 genotypes from the USA and 11 environments were used.
The bilineal AMMI model is represented by:
And the lineal-bilineal SREG model is represented by:
where Yij is the seed yield of genotype i in environment j; μj is the mean value in environment j; i = 1 g; j = 1 .e, g and e being the number of genotypes and environments respectively; and t is the number of principal components (PC) used or retained in the model, with t ≤ min(e,g-1); gi is the effect of the genotype; aj is the effect of the environment; λk (λ1 ≥ λ2 ≥ ≥ λt) are the eigenvalues of each principal component retained in the model. The model is subject to orthonormality constraints on the eigenvectors of the genotypes [αik = (α1k , ,αgk)] and the eigenvectors of the environments [γkj = (γk1, ,γke)], where ∑i α2ik = ∑j γ2kj = 1 and ∑i αik ∑i αik' = ∑j γkj ∑j γkj' = 0 for k ≠ k'; αik y γkj for k = 1, 2, 3 are called primary, secondary, tertiary and so on, effects of genotypes and environments and Eij is the error. The AMMI biplot allows the visualization of the main effects of the genotypes and of the environments, in addition to the most important G x E interactions. The GGE biplot allows visualization of the genotype x environment interaction, and the relationship among genotypes and among environments. The AMMI biplot was constructed with the first PC in the y axis and seed yield in the x axis. The GGE biplot was constructed using the first two PC (PC1 and PC2) derived from subjecting the environment-centered data to singular-value decomposition (Ma et al., 2004). The GGE allows visualization of the which-won-where pattern (i.e., which genotype has the highest yield in which environment), also allows simultaneous visualization of the mean performance and stability of the genotypes, the discriminating ability of a specific environment, and the interrelationship among environments (Yan, 2001; 2002) The analysis was conducted with the SAS system using the programs available at CIMMYT (2004).
RESULTS AND DISCUSSION
There were significant differences among genotypes for biomass yield. The G x E interaction was not significant (Table 4). The highest biomass yield was for the genotypes Rojas, Tape, and Lucero. All of these genotypes are from Argentina, although the genotype Omega-USA did not differ significantly from the last three (Table 5). There were significant differences among genotypes for harvest index, the G x E interaction was not significant (Table 4). The genotype Rahab and Carapé had the highest harvest index of 37.7% (Table 5). This value is considered high for an oilseed crop which levels generally fluctuate between 20 and 30% (Aufhammer et al., 2000). Previous research has indicated that flax has higher harvest index than other oilseed with values approaching 40% (D'Antuono and Rossini, 1995). The 1000-seed weight was not different among genotypes and the G x E interaction was not significant. Test weight was significantly different among genotypes and for the G x E interaction. The highest test weight was for the genotypes Rahab and Nekoma, although test weight can be genetically determined, environmental factors may affect seed filling and ultimately seed weight. Higher than optimum temperatures during seed filling or a severe defoliation may cause a decrease in seed weight (Cross et al., 2003). Of the 11 environments evaluated in this study, we did not observed diseases or insects that caused defoliation, and flowering occurred during the end of November through December when maximum temperatures were below 25 ºC at all environments. Flowering and seed filling was much later in Los Ángeles and Osorno than in Chillán, delaying harvest to February and even March. There was secondary and tertiary flowering on those environments with supplemental irrigation, Chillán and Los Ángeles.
The G x E interaction was significant for seed yield (Table 4). The environments with greater seed yield potential were Chillán 2003-2004 and Los Ángeles 2004-2005. The genotypes Cathay and Prompt had the highest seed yield in Chillán 2003-2004 with 5746 kg ha-1 and 5742 kg ha-1, respectively (Table 6). Genotypes from Argentina had low seed yield the season 2003-2004 due lower stands. The Chillán and Los Ángeles environments had higher seed yield potential because they had supplemental irrigation during flowering and seed filling which promoted secondary and tertiary flowering. Lower seed yield was observed at the Chillán 2004-2005 spring planted environment and Chillán 2006-2007 where bird damage occurred before harvest. Bird damage occurred 2 wk before harvest, finches and pigeons fed on the mature bolls extracting the seeds. Since the G x E interaction was significant the data were analyzed with the AMMI and SREG models to determine the which-won-where for the genotypes and environments.
The AMMI and SREG analyses did not include the Argentinean genotypes, since there were not planted at all environments; however some of the Argentinean genotypes were as high yielding (P < 0.05) as Nekoma and York at Chillán 2003-2004, Los Ángeles 2004-2005, Osorno 2003-2004 and Osorno 2004-2005 (Table 6). Outstanding Argentinean genotypes were Salto, Tape, and Rojas. The Argentinean Omega genotype was always lower yielding at all environments. The general adaptation of Argentinean genotypes to South Central Chile was good and similar to that of North American genotypes.
Visualization of the genotype by environment interaction using the AMMI and SREG models
The information of seed yield contained in Table 6 is graphically displayed in an AMMI biplot (Figure 1). The environment main effect explained 76% of the total variation, the genotypes accounted for only 1% of the variation. The x-axis represents the main effect of genotypes and environments while the y-axis displays the interactions between the genotypes and the environment (Crossa et al., 1991) and reveals that Nekoma' interacted positively with the environments with positive PC1 scores ECHI0304, EOS0607, EOSO0304, EYU0405, ECHI0607, and ECHIPR04. There was little variation among the 10 genotypes in mean seed yield (Figure 1). Mean seed yield across genotypes was 2670 kg ha-1.
Figure 1. Biplot GE, AMMI model for 10 flax genotypes on 11 environments: Chillán 2003-2004 (ECHI0304), 2004-2005 Fall (ECHIOT04), 2004-2005 Spring (ECHIPR04), 2005-2006 (ECHI0506) and 2006-2007 (ECHI0607); Los Ángeles 2004-2005 (ELA0405), Osorno 2003-2004 (EOSO0304), 2004-2005 (EOS0405) and 2006-2007 (EOS0607); and Yungay 2003-2004 (EYU0304) and 2004-2005 (EYU0405). PC 1 = principal component 1. Dotted line indicates mean seed yield.
The 11 environments were divided into four groups. Group 1 includes the environments ECHI0304 and ELA0405 which have the highest mean seed yield. These two environments were irrigated. Group 2 includes environments that were very close to the mean seed yield across all environments and with a PC1 near 0 such as EOS0607, EOSO0304, EYU0405, EYU0304, and EOS0405. The locations Osorno and Yungay are classified in the same agroclimate that may explain why they are in the same group. Group 2 includes ECHI0506 and ECHIOT04 both average yielding environments, but with a negative interaction with the genotype Omega-USA, the lowest yielding genotype at those environments. Group 4 has the two lowest yielding environments ECHI0607 and ECHIPR04 both with the lowest mean seed yield. These two environments were seriously damage by birds before harvest that reduced seed yield. All genotypes were damaged similarly. Genotypes with a PC1 near 0 have a general adaptation to all the environments while those genotypes with PC1 positive or negative are adapted specifically to the environments with PC1 of the same sign either positive or negative (Romagosa and Fox, 1993) (Figure 1). Nekoma' and Prompt' were better adapted to ECH0304, EOS0304, and EOS0607 while Linott', Omega-USA', and Pembina' were better adapted to the environments in Group 3. The other genotypes Cathay, Rahab, Bison, Neche, and York have a general adaptation to all environments, i.e. they had a better yield stability (Diepenbrock et al., 1995).
The GGE biplot allowed us to evaluate the performance of the genotypes in each environment (Figure 2). The which-won-where pattern can be visualized only by the polygon view of a GGE biplot. The polygon is formed through connecting the markers of the four genotypes with the greatest seed yield response to the environments whether positive or negative. This indicates that Nekoma' had the highest seed yield followed by Cathay' and Rahab'. Oppositely, Omega-USA' was the genotype with the lowest seed yield. The ideal genotypes are those that have a high mean yield and a PC2 close to 0, or high stability. Nekoma' has both of these characteristics being the ideal genotype for this analysis. It is interesting to note that Nekoma' biomass yield and harvest index were not among the greatest values. This may indicate that greater biomass yield is not necessarily correlated with seed yield. The ideal environment is also the one with high seed yield and high stability. The only environment that was high seed yield was ECH0304 but its stability was low PC2 = 0.78. The 11 environments divided into three apparent groups in terms of their discrimination of the genotypes. Group 1 includes a single environment, ECHI0304 and Group 2 consists of ECHI0506 andECHIOT04. Group 1 and 2 were negatively correlated as evidenced by the large angle between them. As a result genotypes that yielded well in Group 1 would yield poorly in Group 2 and genotypes that yield well in Group 2 would yield poorly in Group 1. Group 3 includes all the other environments and is located near the origin indicating that this group of environments was not discriminating of the genotypes, i.e. all genotypes had similar seed yield in those environments. The GGE biplot revealed five apparent groups of genotypes in terms of their response to the environments (Figure 2). The first four groups included single genotypes, Nekoma, Cathay, Linott, and Omega-USA. Group 5 consisted of all the other genotypes evaluated.
Figure 2. Biplot GGE for model SREG for 10 flax genotypes on 11 environments: Chillán 2003-2004 (ECHI0304), 2004-2005 Fall (ECHIOT04), 2004-2005 Spring (ECHIPR04), 2005-2006 (ECHI0506) and 2006-2007(ECHI0607); Los Ángeles 2004-2005 (ELA0405), Osorno 2003-2004 (EOSO0304), 2004-2005 (EOS0405) and 2006-2007 (EOS0607), and Yungay 2003-2004 (EYU0304), and 2004-2005 (EYU0405). PC1= Principal component 1; PC2 = Principal component 2. Full-line ovals indicate the groups of environments (3) and dashed-line ovals the groups of genotypes (5).
There was a significant difference for the genotype and G x E interaction for oil content (Table 7). The oil content was lower at Chillán and Osorno in 2006-2007 and the highest oil content was for Chillán 2003-2004, Chillán 2004-2005 spring-planted and Yungay 2003-2004 (Table 7). Environmental conditions such as temperature during seed development, N deficiency, and water availability influence the oil content of flax seed (Adugna and Labuschagne, 2003).
All fatty acid profiles evaluated were different among genotypes. The G x E interaction was significant only for oleic acid (Table 8). Alpha linolenic acid is the most import fatty acid in flax seed. All genotypes had more than 54% of linolenic acid. The genotypes Neche and Nekoma were among the genotypes with highest linolenic acid content. Environmental differences were not significant for fatty acid composition.
Flaxseed adapts to South Central Chile although differences were observed among genotypes for biomass yield, harvest index, test weight, and oil content and composition. Seed yield was above 5700 kg ha-1 for some environments and genotypes, indicating the high seed yield potential of some environments in South Central Chile. According to the AMMI and SREG models the 11 environments were classified into four groups by the AMMI and three groups by the SREG models. Genotypes were classified in five groups in the SREG model; four of those groups were single genotypes. Overall mean seed yield was similar for all genotypes; however the genotype Nekoma was the most stable and higher yielding genotype of the North American genotypes evaluated. The environment with highest yielding potential was Chillán 2003-2004, but had low stability.
The greatest oil contents were observed at the Chillán 2003-2004, 2004-2005 Spring sown, and Yungay 2003-2004 environments, and fluctuated between 420 and 530 g kg-1.The climatic differences among environments did not influence oil composition as was expected based on previous research. Flax could become an excellent alternative for crop rotation in South Central Chile given its great seed yield potential.
Funding for this work was provided by FONDEF-CONICYT project Nº D03-I-1100 and FONTEC-CORFO project Nº 202-3446. The authors acknowledge the valuable collaboration of technicians and students on plot planting and management, and data collection and analysis. Thanks also to Wilson González, Luis Zañartu, Alejandra Villar, and Jorge Campos.
Ackman, R.G. 2002. The gas chromatograph in practical analyses of common and uncommon fatty acids for the 21st century. Analytica Chimica Acta 465:175-192. [ Links ]
Adugna, W., and M.T. Labuschagne. 2003. Association of linseed characters and its variability in different environments. Journal of Agricultural Science 140:285-296. [ Links ]
Aufhammer, W., W. Wagner, H.P. Kaul, and E. Kubler. 2000. Radiation use by oil seed crops-a comparison of winter rape, linseed, and sunflower. Journal of Agronomy and Crop Science 184:277-286. [ Links ]
Berglund, D., and R. Zollinger. 2002. Flax production in North Dakota. Bull. A-1038. 12 p. North Dakota Agricultural Experimental Station, Extension Service North Dakota, Fargo, North Dakota, USA. [ Links ]
Carrasco, P. 1998. Descripciones de Suelos VIII Región. 80 p. Universidad de Concepción, Facultad de Agronomía, Departamento de Suelos, Chillán, Chile. [ Links ]
Casa, R., G. Russell, B. Lo Cascio, and F. Rossini. 1999. Environmental effects on linseed (Linum usitatissimum L.) yield and growth of flax at different stand densities. European Journal of Agronomy 11:267-278. [ Links ]
Charlton, B., and D. Ehrensing. 2001. Fiber and oilseed flax performance annual report. Oregon State University, Corvallis, Oregon, USA. [ Links ]
CIMMYT. 2004. Biometrics and Statistics Unit. Manuals and programs. CIMMYT. [Online] Available at http://www.cimmyt.org/en/about-us/organizational-structure/units-a-labs/crop-research-informatics-laboratory/disciplinary-groups-units/biometrics-and-statistics-unit (accessed September 2010). [ Links ]
Cross, R.H., S.A.B. McKay, A.G. McHughen, and P.C. Bonham-Smith. 2003. Heat-stress effects on reproduction and seed set on Linum usitatissimum L. (flax). Plant Cell and Environment 26:1013-1020. [ Links ]
Crossa, J., and P.L. Cornelius. 1997. Sites regression and shifted multiplicative model clustering of cultivar trials sites under heterogeneity of error variances. Crop Science 37:406-415. [ Links ]
Crossa, J., P.N. Fox, W.H. Pfeiffer, S. Rajaram, and H.G. Gauch, Jr. 1991. AMMI adjustment for statistical analysis of an international wheat yield trial. Theoretical and Applied Genetics 81:27-37. [ Links ]
D'Antuono, L.F., and F. Rossini. 1995. Experimental estimation of linseed (Linum usitatissimum L.) crop parameters. Industrial Crops and Products 3:261-271. [ Links ]
Del Pozo, A., y P. Del Canto. 1999. Áreas agroclimáticas y sistemas productivos en la VI y VIII Regiones. Serie Quilamapu Nº 113. Instituto de Investigaciones Agropecuarias, Centro Regional de Investigación Quilamapu, Chillán, Chile. [ Links ]
DIMEAGRO. 2007. Perfil del mercado del lino 2006. [Online] Gacetilla Informativa del sector agroalimentario. Secretaría de Agricultura, Ganadería, Pesca y Alimentos, República Argentina. Available at http://www.sagpya.mecon.gov.ar/new/0-0/programas/dma/newsletters/nro25/newsletter_lino-01-12-2006.php (accessed January 2008). [ Links ]
Diepenbrock, W.A., J. León, and K. Clasen. 1995. Yielding ability and yield stability of linseed in Central Europe. Agronomy Journal 87:84-88. [ Links ]
Dybing, C.D., P.D. Evenson, and C. Lay. 1988. Relationships among daily flower production, length of the flowering period, and seed yield of flax. Crop Science 28:287-292. [ Links ]
FAO. 2009. FAOSTAT database. Agricultural crops: linseed: area harvested/yield. Available at http://faostat.fao.org/ (accessed September 2010). [ Links ]
Froment, M.A., J. Smith, and K. Freeman. 1999. Influence of environmental and agronomic factors contributing to increased levels of phospholipids in oil from UK linseed Linum usitatissimum. Industrial Crops and Products 10:201-207. [ Links ]
Grady, K., and L. Gilbertson. 2009. South Dakota flax variety evaluations 2009. Bull. ExEx8055. October 2009. Cooperative Extension Service South Dakota State Univ. Brookings, SD. [ Links ]
Grady, K.A., and C.L. Lay. 1994. Registration of Prompt Flax. Crop Science 34:308. [ Links ]
Grady, K.A., C.L. Lay, and M.W. Ferguson. 1987. Registration of Rahab' flax. Crop Science 27:362-363. [ Links ]
Hammond, J.J., J.F. Miller, G.D. Statler, and T.J. Gulya. 1991. Registration of Neche Flax. Crop Science 31:847. [ Links ]
Kempton, R.A. 1984. The use of biplots in interpreting variety by environment interactions. Journal of Agricultural Science 103:123-135. [ Links ]
Ma, B.L., W. Yan, L.M. Dwyer, J. Frégeau-Reid, H.D. Voldeng, Y. Dion, and H. Nass. 2004. Graphic analysis of genotype, environment, nitrogen fertilizer, and their interactions on spring wheat yield. Agronomy Journal 96:169-180. [ Links ]
Madhusudhan, K.T, H.P. Ramesh, T. Ogawa, K. Sasaoka, and N. Singh. 1986. Detoxification of commercial linseed meal for use in broiler rations. Poultry Science 65:164-171. [ Links ]
Miller, J.F., J.J. Hammond, and G.D. Statler. 1988. Registration of four flax germplasm lines. Crop Science 28:579. [ Links ]
Miller, J.F., J.J. Hammond, and G.D. Statler. 1992. Registration of Omega' flax .Crop Science 32:1065. [ Links ]
Novoa, R., S. Villaseca, P. del Canto, J. Rouanet, C. Sierra, y A. Del Pozo. 1989. Mapa agroclimático de Chile. Instituto de Investigaciones Agropecuarias (INIA), Santiago, Chile. [ Links ]
Oomah, B.D., E. Kenaschuk, and G. Mazza. 1996. Phytic acid content of flaxseed as influenced by cultivar, growing season, and location. Journal of Agricultural and Food Chemistry 44:2663-2666. [ Links ]
Romagosa, I., and P.N. Fox. 1993. Genotype x environment interaction and adaptation. p. 374-390. In Hayward, M.D., N.O. Bosemark, and I. Romagosa (eds.) Plant breeding, principles and prospects. Chapman & Hall, London, UK. [ Links ]
SAS Institute. 2007. SAS/STAT® User's guide, Version 9.1. 4th ed. Vol. 1. SAS Institute, Cary, North Carolina, USA. [ Links ]
Steel, R.G.D., J.H. Torrie, and D.A. Dickey. 1996. Principles and procedures of statistics: A biometrical approach. 3nd ed. McGraw-Hill, New York, USA. [ Links ]
Vargas, M., y J. Crossa. 2000. El análisis AMMI y la gráfica del biplot en SAS. Centro Internacional de Mejoramiento de Maíz y Trigo (CIMMYT), México D.F., México. [ Links ]
Yan, W. 2001. GGEBiplot- A Windows applications for graphical analysis of multi-environment trial data and other types of two way data. Agronomy Journal 93:1111-1118. [ Links ]
Yan, W. 2002. Singular-value partition for biplot analysis of multi-environment trial data. Agronomy Journal 94:990-996. [ Links ]
Yan, W., L.A. Hunt, Q. Sheng, and Z. Szlavnics. 2000. Cultivar evaluation and mega-environment investigation based on the GGE biplot. Crop Science 40:597-605. [ Links ]
Yan, W., M.S. Kang, B. Ma, S. Woods, and P.L. Cornelius. 2007. GGE Biplot vs. AMMI analysis of genotype-by-environment data. Crop Science 47:643-653. [ Links ]
Zobel, R.W., M.J. Wright, and H.G. Gauch. 1988. Statistical analysis of a yield trial. Agronomy Journal 80:388-393. [ Links ]
Accepted: 07 July 2009.