Significant annual differences (P<0.01) were observed for all traits, with the 2018 season yielding higher seed and oil yields due to favorable precipitation and temperature conditions. In contrast, the drier and hotter 2017 season negatively affected safflower growth and key yield components, such as plant height and branch number, due to rainfall deficit and heat stress. These environmental variations significantly influenced the performance of both ‘Dinçer’ and ‘Yenice’ cultivars, with 2018 providing better conditions for seed filling and oil accumulation.
Plant height
Plant height was significantly affected by year, cultivar, row spacing, sowing norm and the cultivar x row interaction (P<0.01) (Table 1). Plant height was substantially greater in 2018 (84.9 cm) than in 2017 (64.2 cm) due to more favorable climatic factors
(Montemurro et al., 2007; Fig 1). The Dinçer cultivar was 6.8 cm taller than Yenice. The highest plant height was achieved at the 20 cm row spacing and 60 kg ha
-1 sowing norm, while the shortest was at 40 cm spacing and 20 kg ha
-1 application. This indicates that increased plant density stimulates vertical growth by intensifying competition for sunlight (
Moghaddasi and Omidi, 2015;
Uke et al., 2017).
Branch number
Branch number exhibited significant effects from year, cultivar, row spacing, sowing norm and the cultivar x row and row x norm interactions (P<0.01). The highest average branch number was recorded in 2018 (6.2 pieces), attributable to the high and uniform rainfall distribution during the growing season. The Dinçer variety (5.9 pieces) generally produced more branches (Table 1). Branch number significantly varied inversely with plant density: it increased with wider row spacing (60 cm yielded the highest value) but decreased significantly as the sowing norm was increased (20 kg ha
-1 yielded the lowest). This observed reduction in branching in dense plantings is likely due to reduced light penetration and increased competition (
Sharif and Omidi, 2016), as the plant prioritizes vertical growth over lateral development.
Head number
The head number, a critical yield factor, was significantly influenced by year, sowing norm and cultivar x row interaction (P<0.01) and by cultivar (P<0.05). The average head count in 2018 (14.4) was significantly higher than in 2017 (11.0) due to more favorable ecological conditions, specifically beneficial rainfall and temperature (Table 1), (Fig 1). The Dinçer cultivar produced the maximum number of heads. While the effect of different row spacing was statistically insignificant, the head number exhibited a clear inverse relationship with seeding density. The highest number of heads (13.3 pieces) was obtained at the lowest sowing norm (20 kg ha
-1) and the lowest (12.4 pieces) was obtained at the highest norm (60 kg ha
-1). This density-dependent decrease is consistent with previous reports that increased sowing norm reduces the number of heads due to intensified plant competition (
Moghaddasi and Omidi, 2015;
Uke et al., 2017).
Seed number
The number of seeds per head was significantly influenced by year, cultivar, sowing norm and the two-way interactions row spacing x cultivar and row spacing x sowing norm (P<0.01); row spacing alone was also significant (P<0.05). The highest seed count was recorded in 2018 (25.3 seeds) and the Dinçer cultivar produced the maximum number of seeds per head (24.5 pieces) due to genetic differences (Table 1). A wider 60 cm row spacing maximized seed number (Likely by enhancing branching
via increased red light reception), while increasing the sowing norm significantly reduced the count (lowest at 60 kg ha
-1 with 23.2 seeds), a reduction attributed to high plant density causing a post-flowering water deficit (
Sefaoglu and Ozer, 2022;
Caliþkan and Caliþkan, 2018).
1000 seed weight
The Thousand Seed Weight (TSW), a key yield determinant, was significantly affected by year, cultivar, row spacing, sowing norm and all two- and three-way interactions (Table 2). TSW was substantially higher in 2018 (45.3 g) than in 2017 (41.0 g), which is attributed to sufficient rainfall and suitable temperatures during the critical July 2018 flowering period
(Beyyavas et al., 2011). TSW showed an inverse relationship with both row spacing (highest at the narrowest 20 cm with 44.4 g) and sowing norm. This negative response to high density is likely due to intensified competition slowing metabolic activities during the seed-filling stage
(Zarei et al., 2011; Caliskan and Caliskan, 2018).
Seed yield
Seed yield, which is the cumulative result of various yield components and external factors, was significantly affected by year, variety, row spacing and sowing norm treatments (Table 2), (P<0.01). The effects of all interactions, except row spacing x sowing norm, were insignificant (Table 2). Seed yield exhibited significant genotypic differences (
Bella et al., 2019), with Dinçer (1360.9 kg ha
-1) outperforming Yenice (1190.0 kg ha
-1). Contrary to expectations, the first trial year yielded approximately 170.1 kg ha
-1 higher than the second. This lower yield in the second year is likely attributed to climatic factors
(Vicianova et al., 2020), specifically irregular rainfall and adverse temperatures during critical growth stages (emergence, flowering and grain-filling), negatively impacting key yield components (
Sefaoglu and Ozer, 2022). Seed yield decreased as row spacing increased (1420.6 kg ha
-1 at 20 cm vs. 1160.7 kg ha
-1 at 60 cm), demonstrating that wider spacing leads to increased yield components but a lower overall yield per unit area. Conversely, seed yield increased consistently with the increase in sowing norm, peaking at 1360.2 kg ha
-1 with the 60 kg ha
-1 norm (Fig 2). This finding aligns with studies reporting that higher sowing norms increase yield
(Ahadi et al., 2011), although other research suggests wider spacing enhances light utilization and reduces competition, leading to higher individual plant yield
(Berglund et al., 1998; Zarei et al., 2011).
Seed oil concentration
Seed oil concentration was significantly affected by all treatments, as well as the cultivar x row spacing and row spacing x sowing norm interactions (P<0.01) (Fig 3), (Table 2). Oil content is highly dependent on genotype (
Kose and Bilir, 2017;
Kayin et al., 2024) and climatic conditions, notably temperature during the grain filling period (
Weiss, 2000). The higher crude oil ratio observed in 2018 (Wetter year) is attributed to significantly higher total rainfall in May, June and July compared to 2017 (Table 2). Among row spacings, the highest oil content was obtained at 40 cm. While the oil ratio showed no significant change between the 20 and 40 kg ha
-1 sowing norms, it was unexpectedly higher in the densest planting (60 kg ha
-1). This highlights that environmental factors, particularly adequate moisture during critical stages, often override simple agronomic density effects in determining seed oil quality (Fig 3).
Oil yield
Oil yield was significantly affected by all factors, reflecting the combined influence of oil concentration and seed yield (Table 2), (Fig 3). Due to its superior genetic structure, the Dinçer cultivar produced the highest oil yield. The lower oil yield in the first year is attributed to irregular and insufficient rainfall. Oil yield showed a density-dependent response, with the lowest yield recorded at the widest 60 cm row spacing and the highest (330.5 kg ha
-1) obtained at the narrowest 20 cm spacing. Furthermore, oil yield increased proportionally with the sowing norm, reaching its maximum at 60 kg ha
-1. This pattern indicates that seed yield per unit area is the primary factor driving the final oil yield, outweighing the minor changes in oil concentration.
Principal component anaysis (PCA)
The principal component analysis (PCA) identified two effective components (with eigen values > 1) (Fig 4). These components explained the majority of the total variance in both cultivars: 98.7% in Dinçer (First: 28.3%, Second: 70.4%) and 97.6% in Yenice (First: 22.3%, Second: 75.3%). In the Dinçer cultivar, the treatment combinations of 40 cm row spacing with 20 kg ha
-1 and 40 kg ha
-1 sowing norms clustered together, showing strong correlations with branch number, head number and plant height. In contrast, 40x60 cm and 20x40 cm distances were associated with the major yield components: 1000-seed weight, seed oil concentration, oil yield and seed yield. For the Yenice cultivar, the 20x20 cm and 20x40 cm distances clustered, correlating strongly with a wider range of traits including plant height, seed number, thousand seed weight and head number. Additionally, the highest sowing norms (60x2 kg ha
-1 and 60x4 kg ha
-1) correlated strongly with branch number, while the 20x6 kg ha
-1 distance showed high correlations with seed oil concentration, oil yield and seed yield (Fig 3; Fig 4).