Introduction
Canola (Brassica napus L.) is a Brassicaceae oilseed developed by genetic enhancement of two rapeseed species: Brassicas oleracea and Brassica rapa7. Besides being responsible for 15% of worldwide edible vegetable oil production, it integrates a select feedstock group producing 95% of biodiesel, globally 8. Thus, canola outstands as an alternative to oil-sourced dependence, increasing renewable energy and reducing greenhouse gas emissions.
Despite its potential, canola is not fully consolidated in Brazil since its cultivation is indi cated within 35° and 55° latitudes, in temperate regions. Currently, Brazilian production is concentrated in the south portion of the country, with over 90% of production located in Rio Grande do Sul 3. However, studies have shown that this oilseed could be produced in a tropical area such as Cerrado (15° 35’18” S, 47° 43’57” W at 999 m) 1,13. Addi tionally, canola has the advantage of being an interesting Safrinha crop, where short-cycle annual crops such as soybeans and maize can grow in crop rotation systems. Incorporating canola in the soybean/maize system could also reduce disease incidence and enable winter production of vegetable oils when large areas of cultivation remain unoccupied.
Thus, canola tropicalization in Cerrado would expand oil production for energy purposes and human consumption, generating job positions and reducing deforestation. However, canola cropping in the Cerrado, requires evaluating hybrid agronomic performance, assisting decision making and choosing appropriate management, and technologies and equipment. Therefore, our objective was to evaluate agronomic performance of canola hybrids in the tropical Brazilian Cerrado, selecting promising candidates to contribute to crop expansion.
Materials and methods
Experiment installation and conduction
Eight spring commercial canola hybrids were evaluated: Hyola 50, Hyola 61, Hyola 76, Hyola 433, Hyola 571, Hyola 575, ALHT B4 and Diamond. Experiments I and II were irrigated by a fixed sprinkler, applying a total of 163 and 520 mm water, respectively, throughout the whole experiment. Experiment III received 182 mm of irrigation just before flowering onset.
Experiment I was conducted from June 17 to September 14, 2017, experiment II from April 23 to September 6, and experiment III from April 11 to August 24, 2019, in the exper imental area of Embrapa Cerrados, located at 15° 35’18” S, 47° 43’57” W, at 999 masl. The soil is classified as a Red Latosol with high clay content. According to Köppen & Geiger (1936), the regional climate is tropical with dry winter and rainy summer (Aw), with an average annual temperature of 21 °C, relative humidity of 68% and average annual rainfall of 1,668 mm. Average temperature during experiments was 21.2 °C, with total rainfall of 9.2 mm during experiment I and 403.5 mm during experiments II and III (Figures 1 and 2).
The eight genotypes were distributed in a randomized block design (RBD) with four replications. In experiment I, each plot consisting of 16 lines, 5 m long and spaced 0.17 m, had a population density of 40 plants.m-² and a total area of 656 m². Experiments II and III, consisted of plots with 8 lines, 5 m long, spaced 0.17 m, with a population density of 40 plants.m-² and a total area of 450 m².
Seven agronomic traits were evaluated: i) number of days to flowering onset (NDF, days): number of days from sowing to flowering onset, in which 50% of the plants presented at least one flower; ii ) cycle (CY, days): days from sowing to harvest; iii) pod length (PL, cm), average length of five random pods derived from five randomly chosen plants; iv) pod mass (PM, g): average weigh of the randomly chosen pods with grains; v) number of grains per pod (NGP): average grains from five random pods; vi) mass of 1000 (thousand) grains (M1000G, g): weight of 1000 grains from five plants; vii) grain yield (PROD, kg.ha-1): total grain weight in relation to plot area.
Statistical analysis
After checking normality and homoscedasticity by Shapiro-Wilk and Bartlett tests respectively, data were subjected to ANOVA (p<0.05). Mean comparison was assessed by Tukey test (p<0.05).
The following statistical model was used:
where:
Yijk = effect of observing the hybrid i on repetition j in k environment (experiments);
mk = general environment average in k environment;
Hi = hybrid effect;
Bk = block effect;
Aj = environmental effect;
HAij = interaction effect of hybrid i and environment j;
eijk = random error.
The following genetic parameters were estimated: quadratic genotypic component quadratic component of the hybrid x environment interaction ;residual variance ; genotypic determination coefficient (R²); intraclass correlation (r); genetic variation coefficient (CVg (%)) and CVg/CVe ratio. All analyses were performed with Genes software 4.
Results and discussion
Significant differences were detected for NDF, PL and NGP (Table 1, page 24), indi cating variability between hybrids.
A similar result was previously detected by Zare and Sharafzadeh (2012) for the same traits.
Significant environmental variation for all traits, except PL, evidenced strong environ mental influence. CY, PL and NGP showed significant differences for hybrid-environment interaction (HxE), indicating different environmental performance for all hybrids.
Heritability (H²) constitutes the proportion of genetic variance in total phenotypic variance. It allows estimating genetic gains in selection cycles, and for fixed treatments, as in the present study, it represents a determination coefficient (R²) 14. For NDF, H2 resulted in 0.92, followed by 0.72 for PL, meaning that these traits are less influenced by the envi ronment than the others. Other studies have also reported high R² magnitudes (above 0.89) for NDF 12,15,16. Meanwhile, NGP and M1000G presented intermediate R² (0.68 and 0.57, respectively), as previously found 9,11. The lowest R² values detected in this study (less than 0.50), indicated a stronger environmental influence on those traits.
Regarding PL (Table 2, page 24), ALHT B4 and Diamond hybrids did not significantly vary among environments.
As for NGP, three of the eight genotypes (Hyola 50, Hyola 433 and Diamond) had significant environmental variation.
When analyzing within each environment, Diamond hybrid presented the shortest cycle in the three experiments (ranging from 102 to 114 days), being, however, not statistically different from the rest in Experiment I. In general, considering that canola varies from 107 to 166 days 6 these evaluated genotypes can be classified as short-cycle hybrids. Precocity may be associated with climatic conditions, such as high temperature (approximately 32.5 °C in experiment I and 33.5 °C in experiments II and III) and long photoperiod. Canola hybrids with a shorter cycle are vital for incorporating canola in crop rotation systems in Brazilian Cerrado, allowing the species to benefit from the end of rainy periods and reducing mech anized irrigation.
The NGP significantly varied across all three experiments, with the hybrids Diamond, Hyola 50, ALHT B4, Hyola 433, Hyola 76 and Hyola 61 exhibiting the highest number of grains. For this trait, Diamond presented the higher values in experiments II and III, agreeing with the 17-24 grains per pod, previously observed in different canola genotypes 5,15.
Grain yield resulted promising (Figure 3) since all averages resulted higher (above 1,500 kg.ha-1) than the national grain yield average (1,429 kg.ha-1) 2, expecting Hyola 571, Hyola 433, ALHT B4 and Hyola 50 hybrids, that resulted below the national average in Experiment III.
The Hyola 61 hybrid showed the highest grain yield (2,926 kg.ha-1), followed by Diamond hybrid (2,816 kg.ha-1) in experiment I. Hyola 61 had previously exhibited high grain yield stability under both water deficit and intense cold 6. Diamond also outstood in exper iment II for grain yield (3,356 kg.ha-1), surpassed only by Hyola 76 (4,179 kg.ha-1) and Hyola 571 (3,930 kg.ha-1). In experiment III, Diamond (1,977 kg.ha-1) showed the highest grain yield, followed by Hyola 61 (1,821 kg.ha-1) and Hyola 76 (1,637 kg.ha-1).
Conclusion
In summary, observing average estimates, we conclude that Hyola 61 and Hyola 76 outperformed with higher grain yield, followed by Diamond, with high production and shorter cycle. Due to these two key characteristics, Diamond can be considered the best option for crop rotation systems. Therefore, both hybrids constitute valuable alternatives for the Brazilian Cerrado.