Basil (Ocimum basilicum) Landraces Can Be Used in a Water-Limited Environment

Basil (Ocimum basilicum L.) is a member of the Labiatae family and is one of the most widely consumed aromatic and medicinal plants in many countries due to its numerous properties and uses. The objective of the study was to determine whether landraces are better adapted to water-limited environments compared to commercial cultivars. Irrigation levels and genotypes affected plant height and leaf area index, with 25% and 33% higher values observed under complete irrigation, respectively. Additionally, limited water availability resulted in a 20% reduction in dry matter yield and a 21% reduction in essential oil yield over the three years in all of the genotypes tested, specifically in the lower irrigation treatment (d40), compared to the control treatment (d100). The landraces that performed the best under limited water supply were Athos white spike (AWS) and Gigas white spike (GWS), indicating their suitability for environments with limited water resources. The results demonstrate that there are landraces that can be utilized in dryland climates with appropriate water management, enabling water conservation and utilization of fields in water-scarce areas for irrigation purposes.


Introduction
The modern agricultural system used worldwide, based on intensive agriculture, poses a significant threat to biodiversity [1]. The widespread adoption of intensive agriculture practices has resulted in a decrease in the number of species used in agricultural ecosystems, a decline in the use of crop rotation, and the displacement of locally adapted varieties (landraces) [2,3] with a limited number of hybrids and improved varieties [4]. This decline in agrobiodiversity is exemplified by the fact that the majority of the world's population relies on products derived from just 12 crops, namely wheat (Triticum aestivum L.), rice (Oryza sativa L.), maize (Zea mays L.), and potato (Solanum tuberosum L.), which account for 60% of their diet [1]. Furthermore, the consequences of intensive agriculture's impact on genetic erosion are evident in the dramatic reduction of varieties within important crop species. For example, in Indonesia, 74% of rice varieties originated from a single genotype, while in the USA, 50% of wheat varieties used by farmers were derived from only nine varieties. Similarly, 75% of potatoes were sourced from four varieties, and 50% of soybeans came from just six varieties [5].
The implementation of intensive agriculture has resulted in several adverse consequences, including increased soil salinity and degradation of the plant growth environment [6][7][8][9][10][11][12][13]. This degradation is not solely attributed to intensive agriculture but is also influenced by climate change, such as rising temperatures and alterations in rainfall patterns, which will further amplify the vulnerability of agricultural production in the future [14,15]. In order to address these conditions, it is crucial to utilize varieties that are well adapted to existing and anticipated adverse changes [16].
Landraces, with their historical evolutionary trajectory and adaptation to stress conditions, can play a significant role in the development of improved varieties that require to the third measurement (Table 1). Among the genotypes, the tallest variety was Kassandra red spike (KRS), followed by Sweet (a commercial cultivar), and then by Pink spike (PS) cultivar. The shortest variety was Corymb white, followed by the landrace Gigas white spike (Table 1). Significant differences were observed between the two irrigation levels, with the tallest plants found at d 100 , which was by an average 25% higher than at d 40 across all genotypes and growth stages over the three years of the experiment (Table 1). The greatest difference was observed in KRS and the commercial cultivar during the third year of the experiment, where the difference between the d 40 and d 100 treatments ranged from 15% to 102%. PS exhibited the least response to reduced water availability compared to other genotypes. Table 1. Combined effect of genotype (A), irrigation (B) and growth stages (C) on plant height at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IR n ) are presented as d 40 and d 100 , respectively, for the three years of experimentation (2017, 2018, and 2020 Ϯ Means in the same genotype followed by the same letter do not differ significan two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treat stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in tionally, there was an increase in dry weight for most of the genotypes t three-year period, specifically from the initiation of flowering to full bloom dry weight was higher in most genotypes during the year 2017 compared to years. Furthermore, variations were observed among the different genotyp two irrigation treatments. Conversely, in 2018, the dry weight of most cultiv under reduced irrigation, and it was not significantly different between ments. For the landrace Corymb violet (CV), there was an increase in dry w first stage to the second, followed by a decrease from the second to the third for the landrace Athos white spike (AWS), there was also an increase in d tween the first and second growth stages, but no significant difference wa tween the second and third growth stages. Overall, there was a 20% increase in the d100 treatment compared to the d40 treatment over the three growing Table 6. Combined effect of genotype (Α), irrigation (Β) and growth stages (C) on d two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) d40 and d100, respectively, for the three years of experimentation (2017,2018 Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d100 treatment compared to the d40 treatment over the three growing seasons. Table 6. Combined effect of genotype (Α), irrigation (Β) and growth stages (C) on dry weight at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) are presented as d40 and d100, respectively, for the three years of experimentation (2017, 2018, and 2020). Data pre-Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d 40 and d 100 according to the LSD test (p = 0.05).

Leaf Area Index
The leaf area index (LAI) was found to be influenced by growth stages, irrigation level, and genotype ( Table 2). The LAI increased from the first to the second measurement in all treatments and genotypes, but it showed a decrease from the second to the third measurement in most cases. Among the genotypes, Sweet, Genovese, and KRS had the highest LAI values, as indicated in Table 2. Across the majority of genotypes, there were significant differences in LAI between the two irrigation levels, with the highest LAI observed at d 100 , averaging a 33% increase. During the second year, the differences between the two irrigation treatments were smaller, whereas in the third year, the differences were more pronounced in the d 100 treatment. The availability of water had a significant impact on LAI, with limited water resulting in lower LAI in most genotypes, particularly in GWS. Comparing the reduced irrigation treatment to the control treatment, the LAI was considerably lower, as shown in Table 2. The landrace CW did not show any significant changes in LAI. Table 2. Combined effect of genotype (A), irrigation (B) and growth stages (C) on leaf area index (LAI) at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IR n ) are presented as d 40 and d 100 , respectively, for the three years of experimentation (2017, 2018, and 2020

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatm stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in T tionally, there was an increase in dry weight for most of the genotypes tes three-year period, specifically from the initiation of flowering to full bloom dry weight was higher in most genotypes during the year 2017 compared to t years. Furthermore, variations were observed among the different genotypes two irrigation treatments. Conversely, in 2018, the dry weight of most cultivar under reduced irrigation, and it was not significantly different between th ments. For the landrace Corymb violet (CV), there was an increase in dry we first stage to the second, followed by a decrease from the second to the third s for the landrace Athos white spike (AWS), there was also an increase in dr tween the first and second growth stages, but no significant difference was tween the second and third growth stages. Overall, there was a 20% increase i in the d100 treatment compared to the d40 treatment over the three growing se Table 6. Combined effect of genotype (Α), irrigation (Β) and growth stages (C) on dry two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) ar d40 and d100, respectively, for the three years of experimentation (2017,2018 Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d100 treatment compared to the d40 treatment over the three growing seasons.

Chlorophyll Content
The chlorophyll content was observed to be lower in the control treatments compared to the water-limited treatments. Specifically, the CW landrace exhibited a significantly higher decrease in chlorophyll content under the water-limited treatments compared to the other genotypes, as presented in Table 3. On the other hand, certain genotypes such as Athos White Spike (AWS) and Genovese (G) demonstrated an increase in chlorophyll content under the control treatments when compared to the d 40 treatment. However, overall, there was not a substantial difference observed between the two irrigation levels throughout the three years of the experiments. Table 3. Combined effect of genotype (A), irrigation (B) and growth stages (C) on leaf chlorophyll content at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IR n ) are presented as d 40  Means in the same genotype followed by the same letter do not differ significantl two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatm stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in T tionally, there was an increase in dry weight for most of the genotypes tes three-year period, specifically from the initiation of flowering to full bloom dry weight was higher in most genotypes during the year 2017 compared to t years. Furthermore, variations were observed among the different genotypes two irrigation treatments. Conversely, in 2018, the dry weight of most cultivar under reduced irrigation, and it was not significantly different between th 36

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d100 treatment compared to the d40 treatment over the three growing seasons. Table 6. Combined effect of genotype (Α), irrigation (Β) and growth stages (C) on dry weight at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) are presented as d40 and d100, respectively, for the three years of experimentation (2017, 2018, and 2020). Data presented are mean values, where LSD is the least significant difference at 0.05 significance level.

Chlorophyll Fluorescence
The quantum efficiency of photosystem (PS) II was found to be influenced by the growth stage, growing season, genotype, and irrigation treatments. Variations were observed among the different genotypes under the same treatment, and in some cases, differences were also observed between the growth stages and irrigation treatments, as presented in Table 4. The most significant difference was observed in the GWS landrace, where there was a 22% variation between the two irrigation treatments in the third year. No significant differences were observed in the other genotypes. Overall, there was a 6% increase in the quantum efficiency of PS II in the d 100 treatment compared to the d 40 treatment during the third year. Table 4. Combined effect of genotype (A), irrigation (B) and growth stages (C) on chlorophyll fluorescence at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IR n ) are presented as d 40 and d 100 , respectively, for the three years of experimentation (2017, 2018, and 2020). Data presented are mean values, where LSD is the least significant difference at 0.05 significance level.

Year
Chlorophyll Means in the same genotype followed by the same letter do not differ significantl two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatm stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in T tionally, there was an increase in dry weight for most of the genotypes tes three-year period, specifically from the initiation of flowering to full bloom dry weight was higher in most genotypes during the year 2017 compared to t years. Furthermore, variations were observed among the different genotypes two irrigation treatments. Conversely, in 2018, the dry weight of most cultivar under reduced irrigation, and it was not significantly different between th ments. For the landrace Corymb violet (CV), there was an increase in dry we first stage to the second, followed by a decrease from the second to the third s for the landrace Athos white spike (AWS), there was also an increase in dr tween the first and second growth stages, but no significant difference was tween the second and third growth stages. Overall, there was a 20% increase i in the d100 treatment compared to the d40 treatment over the three growing se Ϯ Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d100 treatment compared to the d40 treatment over the three growing seasons. Table 6. Combined effect of genotype (Α), irrigation (Β) and growth stages (C) on dry weight at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) are presented as d40 and d100, respectively, for the three years of experimentation (2017, 2018, and 2020). Data presented are mean values, where LSD is the least significant difference at 0.05 significance level. Ϯ Means in the same genotype followed by the same letter do not differ significantly between the Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d 40 and d 100 according to the LSD test (p = 0.05).

Assimilation Rate CO 2
The assimilation rate (A) was found to be affected by the growth stages, irrigation level, and genotype, as presented in Table 5. The PS and Sweet genotypes exhibited the highest A values, as indicated in Table 5. In certain genotypes, the assimilation rate increased during the d 100 irrigation treatment and declined during the d 40 treatment. Overall, over the three-year study period, the assimilation rate was 15% higher in the d 100 treatment compared to the d 40 treatment. Additionally, in most cases, A was higher during the first growth stage and lower during the second and third growth stages. The greatest difference between the d 100 and d 40 treatments was observed during the second growth stage, where it reached 19%. Table 5. Combined effect of genotype (A), irrigation (B) and growth stages (C) on assimilation rate (A) of CO 2 at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IR n ) are presented as d 40 and d 100 , respectively, for the three years of experimentation (2017, 2018, and 2020). Data presented are mean values, where LSD is the least significant difference at 0.05 significance level.

Year CO 2 Assimilation Rate (A)
Beginning Means in the same genotype followed by the same letter do not differ significantl two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatm stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in T tionally, there was an increase in dry weight for most of the genotypes tes three-year period, specifically from the initiation of flowering to full bloom dry weight was higher in most genotypes during the year 2017 compared to t years. Furthermore, variations were observed among the different genotypes two irrigation treatments. Conversely, in 2018, the dry weight of most cultivar under reduced irrigation, and it was not significantly different between th ments. For the landrace Corymb violet (CV), there was an increase in dry we first stage to the second, followed by a decrease from the second to the third s for the landrace Athos white spike (AWS), there was also an increase in dr tween the first and second growth stages, but no significant difference was tween the second and third growth stages. Overall, there was a 20% increase i in the d100 treatment compared to the d40 treatment over the three growing se Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d 40 and d 100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d 100 treatment compared to the d 40 treatment over the three growing seasons. Table 6. Combined effect of genotype (A), irrigation (B) and growth stages (C) on dry weight at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IR n ) are presented as d 40

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatm stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Ta tionally, there was an increase in dry weight for most of the genotypes tes three-year period, specifically from the initiation of flowering to full bloom. dry weight was higher in most genotypes during the year 2017 compared to t years. Furthermore, variations were observed among the different genotypes two irrigation treatments. Conversely, in 2018, the dry weight of most cultivar under reduced irrigation, and it was not significantly different between th ments. For the landrace Corymb violet (CV), there was an increase in dry wei first stage to the second, followed by a decrease from the second to the third s for the landrace Athos white spike (AWS), there was also an increase in dr tween the first and second growth stages, but no significant difference was tween the second and third growth stages. Overall, there was a 20% increase i in the d100 treatment compared to the d40 treatment over the three growing se Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d 40 and d 100 according to the LSD test (p = 0.05).

Essential Oil Content
One crucial quality characteristic of aromatic and medicinal plants is the essential oil content, which was observed to be influenced by growth stages, year, genotype, and their interactions, as depicted in Table 7. Among the genotypes, Corymb White (CW), Pink Spike (PS), and Kassandra Red Spike (KRS) exhibited the highest essential oil content, whereas the Genovese (G) cultivar had the lowest content (Table 7). In most genotypes, there was an increase in essential oil content with irrigation, as it was higher in the d 100 treatment compared to the d 40 treatment. Table 7. Combined effect of genotype (A), irrigation (B) and growth stages (C) on essential oil content during 2017, 2018, and 2020, at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) are presented as d 40 and d 100 , respectively, for the three years of experimentation (2017, 2018, and 2020). Data presented are mean values, where LSD is the least significant difference at 0.05 significance level.

Year
Essential Oil Content (mL/100 g Dry Weight) Beginning

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatm stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in T tionally, there was an increase in dry weight for most of the genotypes tes three-year period, specifically from the initiation of flowering to full bloom dry weight was higher in most genotypes during the year 2017 compared to t years. Furthermore, variations were observed among the different genotypes two irrigation treatments. Conversely, in 2018, the dry weight of most cultivar under reduced irrigation, and it was not significantly different between th ments. For the landrace Corymb violet (CV), there was an increase in dry we first stage to the second, followed by a decrease from the second to the third s for the landrace Athos white spike (AWS), there was also an increase in dr tween the first and second growth stages, but no significant difference was tween the second and third growth stages. Overall, there was a 20% increase i in the d100 treatment compared to the d40 treatment over the three growing se Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d100 treatment compared to the d40 treatment over the three growing seasons. Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d 40 and d 100 according to the LSD test (p = 0.05).

Essential Oil Yield (mL/m 2 )
The essential oil yield of aromatic and medicinal plants, an important yield component, was observed to be influenced by growth stages, year, irrigation level, genotype, and their interactions, as presented in Table 8. The essential oil yield was found to be 21% higher in the d 100 treatment compared to the d 40 treatment, attributed to the higher biomass under this treatment. Significant variations in essential oil yield were also observed among the different genotypes tested, with Gigas white spike (GWS) exhibiting the highest yield due to its higher biomass, while the Corymb White (CW) and Corymb violet (CV) genotypes exhibited the lowest yield. Means in the same genotype followed by the same letter do not differ significantl two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatm stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in T tionally, there was an increase in dry weight for most of the genotypes tes three-year period, specifically from the initiation of flowering to full bloom dry weight was higher in most genotypes during the year 2017 compared to t years. Furthermore, variations were observed among the different genotypes two irrigation treatments. Conversely, in 2018, the dry weight of most cultivar 4 Ϯ Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d100 treatment compared to the d40 treatment over the three growing seasons. Table 6. Combined effect of genotype (Α), irrigation (Β) and growth stages (C) on dry weight at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) are presented as d40 and d100, respectively, for the three years of experimentation (2017, 2018, and 2020). Data presented are mean values, where LSD is the least significant difference at 0.05 significance level.

Water Use Efficiency
The water use efficiency (WUE) of the plants exhibited variations based on the growth stage, genotype, and irrigation level, as indicated in Table 9. Generally, WUE was higher during the first growth stage and decreased as the plants grew and produced more biomass. Differences were observed among the different landraces as well as between the landraces and the commercial cultivars that were tested. WUE was generally higher in the reduced irrigation treatment compared to the d 100 irrigation treatment. This is because the relationship between water supply and biomass production is not linear, and providing more water can lead to a slower increase in biomass production. The genotypes with the highest WUE under reduced irrigation were Corymb violet (CV) and Genovese, while the landrace with the lowest WUE was Corymb White (CW) ( Table 9). Table 9. Combined effect of genotype (A), irrigation (B) and growth stages (C) on water use efficiency at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) are presented as d 40 and d 100 , respectively, for the three years of experimentation (2017, 2018, and 2020). Data presented are mean values, where LSD is the least significant difference at 0.05 significance level.

Year
Water Use Efficiency Means in the same genotype followed by the same letter do not differ significantl two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatm stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in T tionally, there was an increase in dry weight for most of the genotypes tes three-year period, specifically from the initiation of flowering to full bloom dry weight was higher in most genotypes during the year 2017 compared to t years. Furthermore, variations were observed among the different genotypes two irrigation treatments. Conversely, in 2018, the dry weight of most cultivar under reduced irrigation, and it was not significantly different between th ments. For the landrace Corymb violet (CV), there was an increase in dry we first stage to the second, followed by a decrease from the second to the third s for the landrace Athos white spike (AWS), there was also an increase in dr tween the first and second growth stages, but no significant difference was tween the second and third growth stages. Overall, there was a 20% increase i in the d100 treatment compared to the d40 treatment over the three growing se 11  Ϯ Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d100 treatment compared to the d40 treatment over the three growing seasons. Table 6. Combined effect of genotype (Α), irrigation (Β) and growth stages (C) on dry weight at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) are presented as d40 and d100, respectively, for the three years of experimentation (2017, 2018, and 2020). Data presented are mean values, where LSD is the least significant difference at 0.05 significance level. Ϯ Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Year
Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d 40 and d 100 according to the LSD test (p = 0.05).

Principal Component Analysis (PCA)
The principal component analysis (PCA) was conducted separately for each year of experimentation, considering all cultivars tested and the mean values of the measured parameters under the two different water treatments (d 40 and d 100 ). This analysis provides insights into the relationship between different plant parameters and their contribution to the overall variability observed in the data. The first component represents a set of correlated characteristics related to plant growth and productivity, while the second component captures the variation in physiological traits associated with photosynthetic efficiency and water use efficiency.
More specifically, in the first year (2017), the PCA identified two major components with eigenvalues greater than 1, which together explained 62.67% of the total variability in the data. The first component accounted for 33.82% of the variability, while the second component explained 28.85% of the variability (Figure 1). In terms of plant parameters, plant height, leaf area index, dry weight, and essential oil yield showed a positive correlation with the first component. On the other hand, essential oil content exhibited a significantly negative correlation with the first component. The second component showed a positive correlation with physiological plant characteristics, including chlorophyll content, chlorophyll fluorescence, assimilation rate of CO 2 , and water use efficiency (Table 10).   In the second year (2018), the PCA revealed two major components with eigenvalues higher than 1, which accounted for a total variability of 52.32% in the data.    Similarly, in the third year (2020), the PCA identified two major components with eigenvalues higher than 1, explaining a total variability of 62.89% in the data. The first component accounted for 32.29% of the variability, while the second component explained 30.6% of the variability (Figure 3). The correlation patterns between variables in the third year were also consistent with those observed in the first and second year. The first component showed positive correlations with morphological parameters (plant height, leaf area index, dry weight, and essential oil yield), while the second component exhibited positive correlations with physiological parameters, especially with chlorophyll content and chlorophyll fluorescence (Table 10).

Discussion
To address the impact of climate change, sustainable intensification of agricultural production can be employed, focusing on the efficient utilization of water resources and the development of cultivars or utilization of genotypes that are well-suited to changing environmental conditions. This approach becomes particularly crucial in regions with low rainfall, such as the Mediterranean area, where water stress is a significant environmental factor [33]. Additionally, approximately 18% of the global land surface is situated in semiarid areas characterized by limited water availability, unpredictable rainfall patterns, land degradation, and limited access to modern agricultural technologies. In such regions, research on crop species and cultivars that are resilient and adapted to these challenging environments becomes a critical option [34].

Plant Height
The plant height of basil has been demonstrated to be influenced by soil water availability [26]. In this study, both basil landraces and the commercial cultivars tested exhibited lower plant height when subjected to low water availability. The factors investigated in the present study included genotypes, growth stages, and the year of experimentation. The environmental conditions, as revealed in this study, can impact the plant height of basil plants. Furthermore, significant differences were observed among the three years of experimentation, which could be attributed to variations in weather conditions. Specifically, higher mean temperatures and lower rainfall in certain years, particularly in June and July, may have affected the growth of basil plants [26,27]. This effect was also noted in the evaluation of five commercial cultivars over a two-year period [26,27]. Based on the scatterplots shown in Figure 1 for the first year, it can be observed that Pink Spike and Sweet cultivars exhibited high tolerance to water deficit treatment, as indicated by their high values, particularly along the first component related to morphological parameters and yield. Conversely, Corymb White showed low mean values on both axes, indicating its high susceptibility to water stress. Gigas White Spike displayed a wide range of values, particularly along the second component related to physiological characteristics. Notably, in the d 40

Discussion
To address the impact of climate change, sustainable intensification of agricultural production can be employed, focusing on the efficient utilization of water resources and the development of cultivars or utilization of genotypes that are well-suited to changing environmental conditions. This approach becomes particularly crucial in regions with low rainfall, such as the Mediterranean area, where water stress is a significant environmental factor [33]. Additionally, approximately 18% of the global land surface is situated in semiarid areas characterized by limited water availability, unpredictable rainfall patterns, land degradation, and limited access to modern agricultural technologies. In such regions, research on crop species and cultivars that are resilient and adapted to these challenging environments becomes a critical option [34].

Plant Height
The plant height of basil has been demonstrated to be influenced by soil water availability [26]. In this study, both basil landraces and the commercial cultivars tested exhibited lower plant height when subjected to low water availability. The factors investigated in the present study included genotypes, growth stages, and the year of experimentation. The environmental conditions, as revealed in this study, can impact the plant height of basil plants. Furthermore, significant differences were observed among the three years of experimentation, which could be attributed to variations in weather conditions. Specifically, higher mean temperatures and lower rainfall in certain years, particularly in June and July, may have affected the growth of basil plants [26,27]. This effect was also noted in the evaluation of five commercial cultivars over a two-year period [26,27].
Overall, the plants exhibited a shorter height, on average by 25%, under the d 40 irrigation treatment compared to the control treatment. Similar observations have been reported in other studies involving aromatic and medicinal plants, including basil, using commercial cultivars [26,27,[35][36][37][38][39]. However, the magnitude of this effect can vary depending on the genotype, as evidenced in the present study. The plant height that was found in this study was higher than in other studies involving basil landraces [38,40], which could be attributed to better growth conditions and the inclusion of genotypes with higher growth potential.

Leaf Area Index
Leaf area is a crucial plant characteristic, particularly for plant species where leaves are the primary harvested plant part, such as many aromatic and medicinal plants [26,41]. Limited water availability has been shown to decrease leaf area, consequently leading to a reduction in photosynthesis, dry matter production, and essential oil production [26]. In this study, the leaf area index was reduced by an average of 33% under conditions of low water availability, reaching up to 91% in certain genotypes, such as Pink Spike (PS) during the first year and at the first measurement. Additionally, the growth of the plants and the leaf area index were influenced by weather conditions, resulting in variations between the three years of the experiment. Similar findings were reported in studies involving commercial cultivars, where the most significant reduction in leaf area index occurred under the lowest water treatment. It has also been documented that water availability, environmental conditions, genotype, and growth stages collectively impact the leaf area of basil plants [27].

Chlorophyll Content
Chlorophyll meters, such as the SPAD meter, have been utilized in various studies to assess the impact of abiotic stresses such as nitrogen deficiency and water stress on plant growth and yield [26,27,42]. In the case of basil, the measurement of chlorophyll content under different water availability conditions revealed that it is not influenced by water availability. Therefore, chlorophyll content cannot serve as a reliable criterion for identifying water stress-tolerant genotypes in basil, unlike in some other plant species [43].
Furthermore, chlorophyll meters are employed to determine leaf senescence and investigate how both abiotic and biotic stresses can affect leaf longevity and the production of assimilates [41,42]. These meters provide a means to nondestructively quantify the level of chlorophyll in leaves under a wide range of growth conditions, which is indicative of their photosynthetic capacity and overall physiological status. By monitoring chlorophyll levels, researchers can gain insights into the impact of stress conditions on leaf function and assess the efficiency of assimilate production in plants [41,42].

Chlorophyll Fluorescence
Chlorophyll fluorescence is commonly used to assess the impact of biotic and abiotic stress on PSII photochemistry, particularly the maximum quantum efficiency, in various plant species, including basil [26,27,[43][44][45][46]. The values reported in this study align with findings from other studies conducted on basil and other plant species [27,47,48]. By measuring chlorophyll fluorescence, researchers can evaluate the functionality and performance of the photosynthetic apparatus.
It is worth noting that the photosynthetic electron transport process is generally more resilient to water stress compared to the carbon reduction cycle of photosynthesis. However, under conditions of reduced water availability, there is a decrease in the rate of NADPH generation, which can impact photosynthetic electron transport [44,47]. Furthermore, water stress can lead to an increase in the generation of reactive oxygen species, resulting in photoinhibition and the degradation of key components of the photosynthetic membrane, such as the D1, D2, and CP43 proteins of PSII, as well as pigments and lipids associated with the photosynthetic apparatus [30,49]. These processes can ultimately affect the overall efficiency of photosynthesis in basil and other plant species under water stress conditions.

CO 2 Assimilation Rate
Water stress affects physiological characteristics of most crop plants [50,51]. However, the effect of water stress on aromatic and medicinal plants and especially on different genotypes and landraces was not determined until recently [26,27]. This is particularly important given the growing interest in alternative crop species such as basil.
The reduction in the CO 2 assimilation rate (A) observed in some genotypes of basil under water stress ranged from 4% to 113%, which is consistent with findings in other plant species and commercial cultivars of basil [27]. In particular, landraces exhibited an average increase of 14% in CO 2 assimilation rate under limited water availability, and in some cases, the increase exceeded that of commercial cultivars. When plants experience limited water availability, they tend to close their stomata to conserve water, leading to a reduction in CO 2 assimilation rate (A) [30,48,49]. Prolonged water stress can further exacerbate this reduction in A.
The use of gas exchange measurements to assess the assimilation rate of CO 2 in aromatic and medicinal plants, including basil, has gained attention only recently [27]. It is crucial to employ gas exchange techniques to investigate the impact of abiotic and biotic stress on different genotypes, particularly landraces, compared to commercial cultivars in aromatic and medicinal plants. The effect of water stress on photosynthetic efficiency has been studied in several plant species such as faba bean, wheat, and triticale [52][53][54][55]. Expanding this research to include aromatic and medicinal plants will provide valuable insights into their physiological responses to water stress and help identify genotypes that are better adapted to water-limited conditions.

Dry Weight
Dry weight, as an important agronomic characteristic, has been studied in basil and shown to be influenced by genotype, irrigation treatment, and growing season. While the effect of water stress on dry weight has been determined in recent studies on commercial cultivars, its impact on landraces has not been extensively investigated. These studies have revealed that some landraces exhibit tolerance to limited water availability, while others may be more sensitive compared to the tested commercial cultivars [31,[56][57][58]. This suggests that certain landraces could be utilized in water-limited environments [31,[56][57][58]. It is possible that certain commercial cultivars and landraces have deeper root systems, allowing them to access available water in the soil and meet their water requirements [31,[56][57][58]. Tolerant landraces may also possess other mechanisms that enable them to cope with low water availability, such as completing their life cycle before severe water stress occurs or improving water use efficiency mechanisms [21].
Aromatic and medicinal plants have gained increasing interest as alternative crops that contribute to biodiversity enhancement in agricultural systems. They require lower inputs of fertilizers, pesticides, and water compared to conventional crops [1,59]. Furthermore, aromatic and medicinal plants are considered more resilient to climate change [1,59]. Basil, in particular, is a versatile plant species that can be grown for culinary uses as well as for its essential oil content [1,59]. However, the lack of breeding programs for developing new basil cultivars adapted to various environmental conditions and the limited availability of certified seeds have led to a reliance on landraces in basil production. Landraces play an important role in providing farmers with options for cultivating basil in diverse environmental conditions where certified cultivars may be scarce [1,59].

Essential Oil Content and Yield
The essential oil content and yield of basil were found to be influenced by various factors, including genotype, growing conditions, and growth stage [60,61]. In the present study, the irrigation level had a significant impact on the essential oil content, consistent with findings in other plant species [26,27,61,62]. Interestingly, the essential oil content values reported in this study were higher compared to those in previous studies, where the range was around 0.25 mL/100 g [26]. This difference could be attributed to the better growing conditions that existed in the present study.
It is worth noting that water availability tends to affect dry matter yield more significantly than essential oil content, as observed in this study and reported in other studies as well [26]. The essential oil content ranged from 0.50 to 1.94 mL/100 g, showing considerable variation among the three years of the experiment and the tested genotypes. Landraces, in particular, exhibit greater variation compared to commercial cultivars, which can result in higher or lower essential oil content under different treatment conditions.
Overall, the essential oil content of basil is influenced by multiple factors, and the variation observed among genotypes and growing conditions highlights the complexity of this trait [26]. Further studies are needed to elucidate the specific mechanisms and interactions that contribute to the variability in essential oil content in basil.

Water Use Efficiency
In the present study conducted on basil, it was observed that the Genovese cultivar and Corymb violet had the highest water use efficiency (WUE), while Corymb white had the lowest WUE. These findings align with the present study, where higher WUE was observed in lower irrigation treatments [26]. WUE in basil is influenced by factors such as irrigation level, year of experimentation, growth stage, and genotype. Identifying genotypes that are tolerant to water stress and can efficiently utilize water resources is crucial, particularly under limited water supply conditions [26,27,43,[63][64][65][66][67][68].
The reported values for WUE in the present study were higher compared to those in other studies. This could be attributed to the cultivation of basil for its leaves, which are harvested at full bloom [63][64][65]. This practice results in lower water consumption, allowing for the conservation of water resources for other crops and the environment [63][64][65]68]. Additionally, previous studies have shown that there is a plateau value in the relationship between water supply and crop production in basil, suggesting that further increases in water supply do not significantly impact dry matter yield [26,27].
Overall, understanding the WUE of basil and its relationship with water supply is important for efficient water management and crop production. Identifying genotypes that exhibit high WUE can contribute to sustainable agricultural practices and the conservation of water resources.

Principal Component Analysis (PCA)
Principal component analysis has been used in different studies to investigate the interrelationships between various traits of plant species [69][70][71]. In the present study, it was found that different genotypes exhibited increased tolerance to limited water availability. Similar results have been reported for other crop species, such as chickpea, where seed yield showed a positive correlation with morphological and agronomic characteristics [70,72]. Furthermore, PCA was employed to provide insights into the genetic diversity based on different traits [70,73]. Several researchers have utilized PCA analysis to classify genotypes and evaluate their response to biotic and abiotic stress [69,70,72,73].
The findings of the present study indicate that consistent relationships existed between the measured variables and their contribution to the overall variability in the data across different years of experimentation. The first component captured variation related to morphological characteristics and productivity, while the second component reflected variation in physiological parameters associated with chlorophyll-related traits. Additionally, the observations of different genotypes under two the irrigation levels highlighted the variability in response to water stress among the genotypes and underscored the varying levels of tolerance or susceptibility to limited water availability throughout the three years of the experiment. Similar differences among growing seasons have also been observed in other crop species [69,[73][74][75][76].

Experimental Site and Crop Management
A 3 year (2017, 2018 and 2020) field experiment was conducted in the University Farm of the School of Agriculture of the Aristotle University of Thessaloniki, which is located in Thermi (40 • 32 9 N 22 • 59 18 E, 0 m)., in Greece. The same field was used over the 3 years, and the soil was clay loam with pH (1:1 H 2 O) 7.77; EC (dS m −1 ) was 1.07, organic matter concentration was 12.40 g kg −1 , with 56 kg ha −1 of N-NO 3 , 24 kg ha −1 of P (Olsen), and 204 kg ha −1 of exchangeable K and CaCO 3 concentration of 1.13 g kg −1 . The soil characteristics were determined according to methods detailed by Sparks et al. [77]. The field was fertilized with 100 and 50 kg ha −1 of N and P, respectively, using ammonium phosphate fertilizer before planting. Weather information was recorded during the growth period with an automatic weather station that was close to the experimental site and reported as mean monthly data for all 3 years.
The weather conditions during the 3 years of the experiment varied considerably. In 2017, there was higher rainfall during June but lower rainfall during July and August compared to the 30-year average for the same period. In contrast, 2018 had lower rainfall during the summer months (June, July, and August) compared to both 2017 and the 30-year average. During the third year, 2020, rainfall was low in June and July but higher in August (Table 11). The mean, minimum, and maximum temperatures were higher during all three years of the experiment compared to the 30-year average. Table 11. The main weather parameters (mean relative humidity (RH mean ), rainfall, maximum (T max ), minimum (T min ), and mean (T mean ) temperature) and reference evapotranspiration (ET o ), for the three years and their comparison with 30-year averages. The weather data were recorded with an automatic weather station close to the experimental site.

Plant Cultivars Used in the Study
Twenty landraces were evaluated under field conditions during 2015 and 2016 for their agronomic characteristics and also for their essential oil content and essential oil yield. From the 20 basil landraces, different basil landraces were used in this study and had differences in earliness, biomass production, and essential oil content. The landraces that were tested are shown in Table 12. Ϯ Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d100 treatment compared to the d40 treatment over the three growing seasons. Table 6. Combined effect of genotype (Α), irrigation (Β) and growth stages (C) on dry weight at the two irrigation levels, where 40% and 100% of the net irrigation requirements (IRn) are presented as d40 and d100, respectively, for the three years of experimentation (2017, 2018, and 2020). Data presented are mean values, where LSD is the least significant difference at 0.05 significance level. Means in the same genotype followed by the same letter do not differ significantly between the two irrigation treatments d40 and d100 according to the LSD test (p = 0.05).

Dry Weight
Dry weight was observed to be influenced by genotype, irrigation treatments, growth stage, and year, with values ranging from 87 to 468 g/m 2 , as presented in Table 6. Additionally, there was an increase in dry weight for most of the genotypes tested over the three-year period, specifically from the initiation of flowering to full bloom. Overall, the dry weight was higher in most genotypes during the year 2017 compared to the other two years. Furthermore, variations were observed among the different genotypes between the two irrigation treatments. Conversely, in 2018, the dry weight of most cultivars was higher under reduced irrigation, and it was not significantly different between the two treatments. For the landrace Corymb violet (CV), there was an increase in dry weight from the first stage to the second, followed by a decrease from the second to the third stage. In 2020, for the landrace Athos white spike (AWS), there was also an increase in dry weight between the first and second growth stages, but no significant difference was observed between the second and third growth stages. Overall, there was a 20% increase in dry weight in the d100 treatment compared to the d40 treatment over the three growing seasons. Where L is landrace, and C commercial cultivar.

Crop Management and Experimental Design
The experimental design that was used was a randomized complete block design (RCBD) in a split-split plot arrangement. The two irrigation treatments were the main plots, the genotypes were the sub-plots, and the repeated measures on the three different growth stages were the sub-sub plots with four replications (blocks) per treatment combination. In addition, there were two strips that corresponded to two irrigation treatments of d 100 and d 40 , and in each strip the genotypes (land races and commercial cultivars) were randomized. Each plot had four rows 5 m long, and each row was 50 cm apart. Basil seeds were sown in a mixture of peat and perlite (9:1) on 4 April 2017, 19 March 2018 and 26 March 2020, and when the basil seedlings were 10 cm in plant height, they were hand-planted to the field on 16 May 2017, 25 April 2018 and 11 May 2020 at a rate of 8 plants m −2 . Weeds were controlled with hand weeding or tilling when it was necessary. Pest control was achieved with one application of Deltamethrin at a rate of 500 mL ha −1 .
The different irrigation treatments were applied as described before [26,27], where there were two treatments of 100% and 40% of the net irrigation requirements (IR n ), and are presented as d 100 and d 40 , respectively. IR n was calculated from the equation: The equation was used before to calculate the amount of water that had to be applied. Therefore, ET c was the crop evapotranspiration, while the effective rainfall was given as P e and was taken into account only when it was higher than 4 mm on any day, and entire rainfall was considered as effective rainfall, CR was the capillary rise from the groundwater table, D p was the deep percolation, and R off was the runoff. In this study, the CR, D p and R off were negligible because (a) there was no shallow water table problem in the experimental area; thus, the CR value was assumed to be zero; (b) D p was not assumed since the amount of irrigation water was equal to the deficit amount in the root zone; and (c) irrigation was performed with drip irrigation, and there was no runoff.
Reference evapotranspiration (ET o ) was based on the FAO Penman-Monteith method [78] and was calculated with the following equation: where ET o is the reference evapotranspiration (mm day −1 ), R n is net radiation at the crop surface (MJ m −2 day −1 ), G is soil heat flux density (MJ m −2 day −1 ), T is mean daily air temperature at 2 m height ( • C), u 2 is wind speed at 2 m height (m s −1 ), e s is saturation vapor pressure (kP a ), e a is actual vapor pressure (kP a ), e s − e a is the saturation vapor pressure deficit (kP a ), ∆ is the slope vapor pressure curve (kP a • C −1 ), and γ is the psychrometric constant (kP a • C −1 ). Crop evapotranspiration (ET c ) was calculated with the following equation: where k c is the crop coefficient.
The following values of crop coefficient (k c ) were used: for the beginning of flowering, 0.9, for full bloom, 1.1, and for the end of flowering, 1.0 [79].
When the plants were transplanted, 30 mm of irrigation water was applied for better establishment of young seedlings. Subsequently, the different levels of irrigation were applied about 40 days after transplantation and when the plants were at the vegetative stage. The irrigation date was the same for the two treatments and was determined when the soil moisture was at 70% of field capacity of the full irrigation treatment (d 100 ) that was considered adequate for plant growth in all growth stages, as described by Kalamartzis [26,27]. The water amount for the deficit irrigation treatment (d 40 ) was determined at 40% of the full irrigation treatment. The water was applied with a drip irrigation system after transplanting, with the drippers spaced at 50 cm intervals; the water supply rate of the drippers was 4 L h −1 . The drip irrigation lines were placed every other row. Volumetric soil water content (θ) was measured with a PR2-type profile probe (Delta-T Devices Ltd., Cambridge, UK) and was used to check the soil water content and to confirm the different water levels. The same irrigation system was extensively used in other experiments [26,27,48].

Sampling
During the growth of the crop, three samplings were conducted at different growth stages covering an area of 1 m 2 . The first sampling was at the beginning of flowering, the second at full bloom, and the third at the end of flowering, starting from the first week of July until the first week of August in all years. The following characteristics were determined: agronomical, morphological, and physiological, water use efficiency (WUE) and essential oil content.

Morphological Characteristics
Two morphological characteristics were determined three times during the growth period of the basil plants for the 3 years that the study lasted. Plant height and leaf area index (LAI) are important characteristics that were determined three times before each sampling by measuring the height of five randomly selected plants per plot from the soil to the top of the plant and obtaining an average value for each plot. LAI was recorded three times before each sampling using the AccuPAR system (model LP-80, PAR/LAI Ceptometer, Decagon Devices, Inc., Pullman, WA, USA). 4.6. Physiological Parameters 4.6.1. Water Use Efficiency The water use efficiency (WUE) for the different cultivars and harvests was determined by dividing the dry weight yield by the total water (rainfall and irrigation) that each treatment received [26,80].

Gas Exchange Measurements
Gas exchange parameters were determined with a portable photosynthesis system (LCi-SD, ADC BioScientific Ltd., Herts, UK) equipped with a square (6.25 cm 2 ) chamber. This device was used for measuring CO 2 assimilation rate (A), transpiration rate (E), stomatal conductance to water vapor (g s ), and intercellular CO 2 concentration (C i ) at the beginning of flowering, full bloom and end of flowering. Measurements were performed on six plants from each plot from 09:00-12:00 in the morning to avoid high vapor-pressure deficits and photoinhibition at midday. Instantaneous water use efficiency (WUE) was obtained by dividing A by stomatal conductance (g s ) [81].

Chlorophyll Fluorescence
The minimum Chl fluorescence (F 0 ) and the maximum Chl fluorescence (F m ) were measured also in situ with the portable Z995 FluorPen PAR (Qubit Biology Inc. Kingston, ON, Canada). For each plot, 10 young fully expanded leaves were used before each sampling. The maximum quantum efficiency of photosystem (PS) II was calculated as F v /F m (F v = F m − F 0 ).

Chlorophyll Content
Chlorophyll content readings (SPAD units) were taken with a hand-held dual-wavelength meter (SPAD 502, Chlorophyll meter, Minolta Camera Co., Ltd., Osaka, Japan). This device calculates the leaf chlorophyll content, with maximum absorption at two different wavelengths (400-500 nm and 600-700 nm) and zero absorption in the near-infrared region. However, the device emits radiation in the red and near infrared region (650 nm and 940 nm) and, combining the absorbance at these two wavelengths, calculates the leaf greenness index, giving the final value. For each plot, 10 young fully expanded leaves were used before each sampling. The instrument stored and automatically averaged these readings to generate one reading per plot.

Dry Biomass and Essential Oil Determination
Dry weight yield was determined by conducting three samplings from an area of 1 m 2 ; 8 plants per sampling were sampled from the inner row and were randomly selected and cut at the ground level. From each sample, the biomass was weighed to obtain the fresh weight (kg ha −1 ) and allowed to dry at room temperature for a week. When the samples reached a constant weight, they were weighed to obtain the final dry weight. The leaves and flowers were separated from stems by hand and weighed. The essential oil content was determined by using 40 g of dry leaf material subjected to water distillation for 3 h, using a Clevenger apparatus (Sigma, London, UK), and the extracted essential oils were stored at −20 • C. The essential oil content of the plants was determined by a volumetric method (ml 100 g −1 ) [82]. The essential oil yield was determined by multiplying the essential oil content by the dry weight.

Statistical Analysis
The data were analyzed with mixed linear models and the ANOVA method according to the model that involved the effects (main and interactions) of three factors: "irrigation levels" × "genotypes" × "growth stages". The experiment was implemented according to a randomized complete block design (RCBD) in a split-split-plot design (irrigation levels × genotypes × growth stages), with four replications (blocks) per treatment combination. The irrigation levels were the main plots, genotypes were the sub-plots, and growth stages were the sub-sub plots [83]. A combined ANOVA over the three factors was performed according to the previously described experimental setup. The least significant difference (LSD) criterion was used to test the differences between treatment means, and the significance level of all hypotheses tested was preset at p < 0.05. All statistical analyses were performed using the SPSS software package (ver. 17, SPSS Inc., Chicago, IL, USA). The statistical analysis (ANOVA) with SPSS was performed within the frame of mixed linear models using an SPSS syntax code developed and programmed by the authors. Pearson correlations were also applied for all plant characteristics that were measured, and a principal component analysis (PCA) was performed to investigate the similarity between the different irrigation treatments within each cultivar and year of experimentation.

Conclusions
The present study described the effect of differences in water availability on different landraces and a commercial cultivar in terms of morphological, physiological, and agronomical characteristics and essential oil content of the different genotypes. It was found that the basil genotypes that were tested were not very sensitive to limited water supply; a reduction of water by 60% compared with full irrigation did not significantly affect the dry weight, as it was lower by 34% compared with the full irrigation. Moreover, it was found that there are some landraces that were not affected by the limited water supply and could be used in areas with low water availability and help to conserve precious water resources.
Author Contributions: All of the authors contributed to the manuscript significantly. I.K. and P.P. conducted the experiments and analysis of results. C.D. was responsible for conducting the experiments and also writing the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.