Heterosis for seed, oil yield and quality of some different hybrids sun ﬂ ower ☆

– Twenty-one hybrids of sun ﬂ ower were produced by crossing 7 introduced cytoplasmic male sterile lines (CMS-lines) with 3 restorer lines (RF-lines) using line (cid:1) tester mating design. The twenty-one hybrids, three restorers, seven maintainer lines (B-lines) were evaluated. The experiment was conducted in a randomized complete block design of three replications. Mean squares due to genotypes, parents (P), crosses (C), lines (L), testers (T), P vs. C, for stearic acid and line (cid:1) tester for palmitic acid. The inbred lines and their F 1 hybrids differed signi ﬁ cantly in their mean values of the traits under study. The variances due to speci ﬁ c combining ability (SCA) were higher than general combining ability (GCA) variances for all the studied traits, showing non-additive type of gene action controlling the traits. Non-additive type of gene action can be utilized for varietal improvement through heterosis breeding. Heterosis values for seed yield plant (cid:3) 1 were positive and highly signi ﬁ cant relative to both the parental mean (17.68 – 72.38%) and the better parent ( (cid:3) 2.86 – 56.842%). Signi ﬁ cantly and negative heterosis was recorded in the case of linoleic acid relative to the parental mean ( (cid:3) 81.24 to (cid:3) 38.02%) and better parent ( (cid:3) 66.24 – 22.87%). With oleic acid, the heterotic effect ranged from (cid:3) 14.18 to 39.59% (parental mean) and from (cid:3) 15.06 to 38.72% (better parent). Therefore, these results are valuable for the improvement of quantitative as well as qualitative traits in sun ﬂ ower breeding material to ful ﬁ ll the edible oil requirements.

Mots clés : Tournesol / rendement en huile / Hétérosis / lignée Â testeur / acides gras 1 Introduction Sunflower (Helianthus annuus L.) is an important edible oilseed crop and ranks fourth in terms of its global edible oil production after palm, soybean and rapeseed. Sunflower seeds contain a high level of oil content (40-50%) (Naeem et al., 2019). In conventional sunflower oil, 90% of the total fatty acids content is comprised of linoleic acid (C-18:2), oleic acids (C-18:1), and 8-10% of mainly palmitic acid (C-16:0) and stearic acid (C-18:0). According to Friedt et al. (1994), in addition to conventional fatty acids, sunflower oil also contains several other fatty acids, but is present only in traces (C14:0, C16:1, C14:1, C20:0, C22:0). Sunflower oil with high oleic acid content is nutritionally similar to olive oil which is considered superior to other types of seed oil (Doty, 1978;Hamed et al., 2020). Grundy (1986) and Soliman et al. (2019) also suggested that a diet rich in monounsaturated fatty acids i. e. oleic acid reduces cholesterol in blood plasma (reducing the risk of coronary heart diseases), has a greater shelf life and a high degree of oxidative stability Zahran and Najafi, 2020;Farrag et al., 2020). The main breeding objective of sunflower is to develop high-yielding, diseaseresistant hybrids with high oil quality (Dudhe et al., 2009).
Sunflower hybrid breeding was started economically in discovering cytoplasmic male sterility CMS by Leclercq (1969) and restorer genes by Kinman (1970). Line Â tester analysis is an extension of this method in which several testers are used (Kempthorne, 1957). Commercial exploitation of heterosis for a particular location (environment) requires isolation of suitable inbred lines and the development of hybrids. To accomplish this task, one has to know the genetic diversity of the available germplasm and the combining ability of the parents. For improving the yield potential of varieties and hybrids, the decision should be made on the choice of the right parent for hybridization. The higher heterosis in hybrids depends on the combining ability of male and female inbreds (Tan, 2010). Duotype sunflower hybrids for higher seed yield and oil traits can be developed with the use of prospective inbred lines, however hybrid superiority over male and female inbreds is an important consideration to evolve thriving F 1 hybrids (Meena et al., 2013). A wide range of heterosis has been reported both for seed yield and oil quality in sunflower by various authors (Joksimovic et al., 2006;Aslam et al., 2010;Chahal et al., 2019).
The predominant role of SCA has been determined for yield and other yield contributing components in sunflower Aleem et al. (2015) while others explained the superior effect of GCA effects over SCA for various traits contributing towards yield (Machikowa et al., 2011). Higher SCA variances as compared to GCA variances were also reported for achene yield per plant, seed and oil yield per hectare (Memon et al., 2014), palmitic, stearic, oleic acid and linoleic acid (Shamshad et al., 2016;Rizwan et al., 2020). Higher GCA variances as compared to SCA variances were also reported for achene yield per plant (Kholghi et al., 2014), palmitic acid, stearic acid and oleic acid (Joksimovic et al., 2006).
The main purpose of this study is to identify superior cross combination for seed and oil yield, as well as for oil quality, which identified as promising crosses and these crosses maybe need further evaluation for commercial exploitation.

Field trial
Seven cytoplasmic male sterile (CMS) lines (A-lines) and three fertility restorer lines (Rf-lines) of sunflower. The A-lines were A 1 , A 5 , A 7 , A 9 , A 11 , A 13 , and A 33 . The tester, Rf-lines, Rf 1 , Rf 9 , and Rf 18 are male restorer lines. The female lines used in the experiment were cytoplasmic male sterile lines (CMS) as a line with their maintainer lines (B lines) and males were restorers (R lines). Parental lines (cms and Rf) in the study are given in Table 1.
Hybrid combinations were created by crossing A-sterile lines with Rf-restorer Research Institute, Agricultural Research Center (ARC), Egypt (22°, 32°N latitude to 24°, 37°E testers during year 2016 at Giza Agricultural Research station, Field Crops longitude). The crossing was undertaken into line Â tester fashion and seeds were harvested separately to study heterosis. The three restorer inbred lines (testers) were crossed with the seven CMS lines during the flowering period. The twenty-one single crosses were obtained by bagging the sterile heads before flowering and the pollen grains were collected from each of the three restorer lines. The stigmas of the seven male sterile lines were pollinated with the collected pollen. The twenty-one obtained sunflower crosses, the three testers, the seven fertile lines (B-lines) were planted at Giza Research Station, Giza Governorate on 30 th July, 2017. A randomized complete block design (RCBD) with three replications was used.
The plot size was 5 rows, 4-meter-long, and 60 cm apart. Planting was done in hills spaced 20 cm apart. Seedlings were thinned to one plant per hill before the first irrigation (two weeks after planting). Per hectare 240 kg superphosphate (15.5% P 2 O 5 ) was added during seedbed preparation, while 120 kg/ha potassium sulfate (48% K 2 O) was added after 15 days from planting. 350 kg/ha ammonium sulfate (20.5% N) was split into two equal doses 15 and 30 days after planting, irrigation was applied every 20 days. Physical and chemical soil analyses of the field experiments (Tab. 2) were performed at laboratories of the Soil and Water Research Institute of ARC, Egypt. The required weather data for the experimental site through the growing season were obtained from Central Lab for Agricultural Climate, Agricultural Research Center at Giza, Governorate, Egypt (Tab. 3).
Data of ten randomly selected plants were recorded from each replication for the characters like seed yield/plant (g), seed yield per hectare (kg), oil yield per hectare (Kg). The heterosis for these traits was estimated according to Wynne et al. (1970) using equations (1) and (2): Better parent heterosis ð Where: -F 1 = single cross hybrid; -MP = related mid-parent; -BP = related better parents.

Oil content (%)
The oil extraction from sunflower seeds was performed in the Fats and Oils Department, National Research Centre, Cairo, Egypt according to the Soxhlet extraction method. The n-hexane was used as extraction solvent at a ratio of 1:10, a sample to solvent. Oil content (%) of seeds was determined according to AOAC (1990).

Fatty acids composition
The fatty acid composition was determined by the conversion of oil to fatty acid methyl esters (FAMEs) according to the modified method by Zahran and Tawfeuk (2019). The FAMEs were separated with an HP 6890 plus gas chromatography (Hewlett Packard, USA), using a capillary column Supelco TM SP-2380 (60 m Â 0.25 mm Â 0.20 mm), (Sigma-Aldrich, USA), Detector (FID) and the injector and detector temperature was 250°C. The column temperature was 140°C (hold for 5 min) and rises to 240°C, at a rate of 4°C/min, and holds at 240°C for 10 min. The carrier gas was helium at a flow rate of 1.2 mL min À1 . FAMEs were identified by comparing their relative and absolute retention times to those authentic standards of FAMEs (Supelco TM 37 component FAME mix). The fatty acid composition was reported as a relative percentage of the total peak area.

Statistical analysis
The line Â tester analysis was calculated according to Kempthorne (1957). The sum of squares for the F 1 single crosses was partitioned into components due to testers (males), lines (females), and line Â tester interaction. The analyses reported in this study were performed with MS-EXCEL (2007) with spreadsheet formula commands.

Analysis of variance
Plant breeders always perform breeding programs aiming to improve and increase the productivity of their plants. In this respect, breeders would try to obtain superior F 1 hybrids or develop new cultivars. Mean squares for line Â tester analysis of lines, testers, and their interactions for all studied traits are presented in Table 4. Results revealed that genotypes and parents exhibited highly significant differences for all studied traits, except parents for palmitic acid indicating that variability existed among all inbred lines. Data revealed that crosses were highly significantly different for all studied traits. Mean squares due to parents vs. crosses were significant for all studied traits. Significant differences in parent's vs. crosses indicated the presence of heterosis in the crosses that may be manifested for the development of high yielding sunflower hybrids. Meanwhile, the lines revealed significant differences for all studied traits. Testers revealed significant differences for all studied traits except stearic acids. Line Â tester interaction exhibited significant differences for all studied traits except palmitic acid. These results showed also great diversity for both lines and testers in seed yield, which contributed to the performance of their respective crosses. The results were following the findings of Manzoor et al. (2016) for seed yield per plant and fatty acids, Harun (2019) for fatty acids, Lakshman et al. (2020) for seed and oil yield per hectare, Haddadan et al. (2020) for seed yield per plant and seed and oil yield per hectare Rizwan et al. (2020) for seed yield per plant and fatty acids.

Mean performance
The mean performance of parents and 21 hybrids for studied traits are presented in (Tab. 5). There were significant differences in seed yield plant À1 , seed yield ha À1 , oil yield ha À1 , linoleic acid, oleic acid, palmitic acid and stearic acid between the A lines and Rf testers on one side and their F 1 hybrids on the other indicate the presence of genetic differences among the studied genotypes. A 5 and A 7 had the highest head diameter, 100-seed weight, and the number of seeds plant À1 , as well as the number of leaves. However, A 1 and A 33 produced higher 100-seed weight and more unfilled seed %. The superiority of A 5 and A 7 in seed yield and most of its components might be attributed to the improved plant type characters; i.e., dry matter production, leaf area index (LAI), and the number of filled seeds that compensated for its lower 100-seed weight.

Heterosis effects
The seed yield of sunflower is a complex character that depends on many traits and varies with the environment. The changeable magnitude of mid parent and better parent heterosis was observed in the hybrids for seed yield per plant. Results in Table 6 showed that seed yield had significant and positive heterosis for all hybrids over mid parent and 16 crosses over better parent respectively. The hybrid A 13 Â Rf 1, A 13 Â Rf 18 , and A 1 Â Rf 9 had the highest significant positive values for heterosis for mid, better parents. The combinations A 5 Â Rf 1 which were best for oil content, 100-seed weight and head diameter, and the number of seeds plant À1 respectively, these combinations involving high Â high parents, indicating that additive Â additive effects which are fixable components. The combinations A 13 Â Rf 1 , A 1 Â Rf 9 , and A 13 Â Rf 18 have low Â high and low Â low combining parents, respectively, indicating good complementation between favorable alleles of the involved parents. Present results are in agreement with those of Hladni et al. (2007) who reported that heterotic values for seed yield were significantly positive relative to parental average as well as better parents. Higher heterosis for seed yield per plant was also reported by Lakshman et al. (2020). The statistical differences within and between parental lines and resulting cross combinations indicate real genetic differences and heterotic impact of high magnitude on the important agronomic traits was expected. A highly significant heterosis  impact on seed yield plant À1 , seed and oil yield ha À1 and fatty acids (up to some extent) both relative to parental mean and better parent was noted.

Seed yield
Significant and desirable positive mid-parent heterosis was observed in all 21 F 1 hybrids, while twelve hybrids manifested heterobeltiosis revealing that hybrids could produce higher seed yield due to dominant or over-dominant genes. The superiority of hybrids over open-pollinated populations in terms of uniformity, productivity, yield stability, oil content, and tolerance to pests and diseases shifted the breeding emphasis from population improvement to heterosis breeding. The hybrids such as A 13 Â Rf 18 , A 1 Â Rf 9 and A 13 Â Rf 1 expressed high relative heterosis of 133.01, 131.25 and 126.24%, and heterobeltiosis of 105.54, 75.07%, and 119.48%, respectively (Tab. 6). The superiority these hybrids might be due to seed yield plant À1 and yield attributes. Positive midparent heterosis has also been reported by Habib et al. (2007) in sunflower. Similarly, positive heterosis over the best parent has also been reported by Memon et al. (2015), Depar et al. (2017) and Khan et al. (2019) and Lakshman et al. (2020) for seed yield (kg/ha À1 .) in sunflower. It is, therefore, concluded that five hybrids viz., A 1 Â Rf 9 , A 13 Â Rf 1 , A 13 Â Rf 18 , A 5 Â Rf 1 , and A 7 Â Rf 18 can be exploited for hybrid seed development on a commercial basis.

Oil yield
Oil yield is an important trait in sunflower which depends on the oil content of the genotype. The results on heterosis for oil yield kg/ha À1 depicted in Table 6 suggested that hybrids A 13 Â Rf 18 (155.96%), A 13 Â Rf 1 (126.44%), A 1 Â Rf 9 (143.79%), and A 5 Â Rf 1 (123.77%), gave high relative heterosis % and better parent heterosis of 135.22, 115.14, 75.08 and 111.08% for oil yield kg/ha À1 . The results showed that oil yield had significant and positive heterosis for all hybrids over a mid and better parent. The superiority of these hybrids in seed and oil yields ha À1 may be due to their genetic constitution and its capability of withstanding climatic fluctuation and soil conditions than ones and related to the increase in root length, a number of leaves plant À1 , leaf area index, head diameter, 100-seed weight, seed weight plant À1 , daily seed weight plant À1 and oil content. The extent of heterotic effects for oil yield kg ha À1 was greater than other yield and oil traits which indicated that oil yield may be improved in further generations through simple selection procedures. Present results are following those of Memon et al. (2015), who observed heterosis over mid parent and a better parent and noted a higher magnitude of average heterosis of (0.08-194.00%), and (À30.93-182.47%) for the better parent in oil yield kg/ha. Lakshman et al. (2020) also observed high positive heterosis over mid and better parents for oil yield.

Linoleic acid (C18:2)
Linoleic acid is an important v-6 (omega-6) fatty acid out of major polyunsaturated fatty acids because it has health benefits of lowering blood cholesterol levels (Orsavova et al., 2015). The heterosis over mid parent for linoleic acid content ranged from À81.24% (A 1 Â Rf 9 ) to À42.81-38.49% (A 7 Â Rf 1 ), (Tab. 7). Linoleic acid belongs to the unsaturated fatty acid group. Similar to oleic acid positive heterosis is considered to be desirable for this trait also. As many as twenty-eight hybrids recorded significant positive heterosis over mid-parent. The heterobeltosis ranged from À66.24% (A 1 Â Rf 9 ) to 22.87% (A 7 Â Rf 18 ). The fatty acid composition changes depending on genotypes and some other factors such as environmental conditions, planting and harvesting time (Roche et al., 2006). The three hybrids recorded significant positive heterosis over the better parent. These results are in agreement with the results reported by Shamshad et al. (2016) and Harun (2019). Linoleic acid constitutes unsaturated fatty acids which are desirable from the health point of view. So significant positive heterosis is desirable for linoleic acid to improve the quality of sunflower oil.

Oleic acid (C18:1)
The high oleic sunflower types are superior to regular sunflower, soybean and peanut oils due to suitability for cooking, and frying for better resistance against heat (Smith et al., 2007). This is an important v-9 (omega-9) fatty acid. The heterosis over mid parent ranged from À14.18% (A 7 Â Rf 18 ) to 39.59% (A 11 Â Rf 1 ), (Tab. 7). Oleic acid is considered to be important from a health point of view, as it belongs to the unsaturated group of fatty acids and enhances the shelf-life of oil due to its oxidative stability. Thus, the emphasis was given to exploit positive heterosis for oleic acid Table 6. Heterosis over mid-parent (MP), better-parent (BP) for seed yield plant À1 , seed and oil yield ha À1 in sunflower genotypes. content in sunflower hybrids. In the present investigation, thirteen of the twenty-one experimental hybrids recorded significant positive heterosis over a mid-parent, whereas, eight hybrids recorded significant positive heterosis over the better parent. Some environmental factors, such as temperature, sunlight and precipitation, affect the growth of sunflowers differently. Every 1°C increase in temperature causes a 2% increase in oleic acidcontent (Demurin et al., 2000). Grunvald et al. (2013) reported that the temperature, especially during the maturation of the seeds, the amount of oleic acid in the oil of conventional sunflower genotypes could exceed 70%.
Higher temperatures led to average increases of up to 35% for this fatty acid. Oil and fatty acid composition in seeds are important targets in sunflower breeding.

Palmitic acid (C16:0)
The range of heterosis over mid parent was from À22.15% (A 11 Â Rf 18 ) to 11.65% (A 5 Â Rf 18 ) and the number of hybrids with negative heterosis was 14, of which, all were significant, the range of heterobeltosis was from À25.92% (A 11 Â Rf 18 ) to 4.81% (A 7 Â Rf 18 ), (Tab. 7). Twenty hybrids recorded significant negative heterosis over the better parent. Heterosis for palmitic acid has been recorded by Tan (2010). As palmitic acid belongs to the unsaturated group of fatty acids, a lot of health risks are involved with a higher concentration of this fatty acid. In the present study hybrids (A 11 Â Rf 18 ) exhibited significantly low levels of palmitic acid and these might serve as potential hybrids, which can be used in the future breeding programs.
3.10 Gene action and heritability 3.10.1 Genetic components for seed yield and quality traits in sunflower Genotypes are presented in Table 8. Results indicated that the non-additive genetic variance including dominance (d 2 D) was larger than their corresponding additive genetic variance (d 2 A) for all studied traits. It was also indicated that nonadditive genetic variances played a major role in the inheritance of these traits. This showed that the hybridization program could be effective in the improvement of those traits. The importance of non-additive variances was verified by the dominance degree ratio, which was more than uniform for all studied traits Table 6. The preponderance of non-additive gene action for these traits was supported in the results of Chahal et al. (2019) and Lakshman et al. (2019). Heritability values in broad and narrow senses were calculated and results are shown in Table 8. Results revealed that broad heritability (H 2 b.s ) estimates were larger than their corresponding values of narrow-sense heritability (H 2 n.s ) for all studied traits. Values of heritability in the broad sense ranged between 94.67 for linoleic acid and 67.91 for seed yield plant À1 , while the heritability in the narrow sense ranged from 6.23 for oleic acid to 10.61 for seed yield plant À1 . These results were in agreement with those obtained by many other authors Attia et al. (2012) and Memon et al. (2014).

Contribution of lines, testers and line Â tester interactions
Lines, testers, and their interaction revealed different contributions in the expression of the studied traits Table 8. The contribution of lines in the expression of seed yield plant À1 (43.10%), seed yield per hectare (40.99%), oil yield per ha À1 (43.66%), palmitic acid (52.09%), stearic acid (65.08%), oleic acid (56.64%), and linoleic acid (56.83%) was the greatest. The contribution of testers in the expression of seed yield plant À1 and seed yield ha À1 and oil yield ha À1 was the greatest. Its contributions in the expression of stearic, oleic, and linoleic acids were almost neglected. Interaction between lines and testers expressed high contributions in many traits, being the highest in oleic acid and linoleic acid.

Conclusion
Significant differences existed among the genotypes (inbreds and hybrids) in their mean values of seed yield and fatty acids traits. The heterosis values for seed yield per plant and fatty acids were highly significant in almost all of the hybrids both mid-parent and a better parent. The results revealed that additive and non-additive gene action was involved in the inheritance of all traits. The dominance variances (s 2 D) were more for seed yield and oil quality as compared to additive variances (s 2 A) and high heritability was found in these traits. This study may prove useful in the future development of high-yielding sunflower hybrids.