Evaluation of Greek Tobacco Varieties (Nicotiana tabacum L.) Grown in Different Regions οf Greece

Abstract

Tobacco (Nicotiana tabacum L.) is an important industrial crop grown in more than one hundred countries worldwide, with high adaptability on a great variety of soils, and under quite diverse climatic conditions. Information regarding the chemical composition of tobacco leaves is important for the tobacco industry because it generally may exhibit important differences among different locations. In this study, five oriental and six flue and air-cured type tobacco varieties were cultivated in five areas of Greece for three cultivation periods, in order to evaluate the effect of these three factors (variety, area, and period), on the yield as well as nicotine content, and sugar and nitrate content of the produced leaves. For the oriental type varieties, the results showed that the area exhibited significant interaction with both variety and the cultivation year, related to the yield. Concerning the nicotine content, no significant differences were observed regarding the areas of experimentation; however, it significantly varied within the varieties. In particular, the nicotine content of Katerini 53 exhibited a significant difference compared to Xanthi 2A, which had the highest content of all tested varieties. Sugar content was affected significantly by all the factors considered, with the area of Xanthi exhibiting lower values compared to all other varieties. Regarding the flue and air-cured varieties, the only interaction that was found to be statistically significant, when the yield was assessed, was between the cultivation year and the area. It was found that there was a significant interaction between cultivation year and area for the nicotine content as well. Furthermore, cultivation year and variety significantly affected the sugar content. The varieties Burley 21E and NC7LC gave the lowest sugars. In general, it was observed that in the case of the oriental type varieties, the variety adaptation in specific regions resulted in higher yields, with the paradigmatic example of the variety Katerini 53 in the Katerini area, and Xanthi 81 in the Xanthi area. On the other hand, in the case of the flue and air-cured varieties, Karditsa was the area that exhibited the lowest yield of all tested varieties.

1. Introduction

Tobacco (Nicotiana tabacum L.) belongs to the family Solanaceae and is one of the most important industrial crops with a worldwide production of 4,900,000 tonnes. Raw tobacco produced in the EU accounts for about 3.7% of the world’s total production, and 7.5% of the world’s marketing. Nowadays, Greece still is in the enviable position of being the world’s second-largest producer of oriental tobacco and belongs to the eight countries that produce 99% of EU tobacco production [1,2]. The structural features of Greek agriculture with the many small size farms, and the surplus of labor with high expertise in the agricultural sector were key factors that reinforced the spread of tobacco growing in the north and central part of the country, where soil and climatic conditions were favorable [3]. However, due to changes in the agriculture sector and raising awareness against smoking, tobacco consumption has declined, both nationally and globally [4].

Tobacco varieties are characterized as sun-cured, flue-cured, air-cured, and fire-cured according to the curing process during which the raw tobacco leaves are dried, lose their green color, and obtain a different color, ranging from yellow to red-brown or black-brown. The main cultivated varieties in Greece are the sun-cured varieties of oriental type, and to a lesser degree, the flue-cured Virginia and light air-cured Burley, which are large-leaf tobaccos that require more inputs [5]. Greek oriental tobacco has high-quality color, aroma, and flavor and has excellent burning properties because of suitable soil characteristics and climate conditions. [6]. Therefore, Greek oriental varieties are used as references in newly introduced varieties and areas with different ecological conditions [7,8,9].

Nicotine is the primary alkaloid in tobacco and is one of the main reasons for its commercial production, while sugars provide mildness when smoking tobacco and had been found to positively affect its quality [10]. Oriental tobacco has sufficient sugar content with less nicotine, as well as a more favorable aroma than the other types. Therefore, oriental tobacco is usually blended with Virginia and Burley tobacco varieties, which exhibit a higher level of nicotine [11] and are used in American blend type cigarettes [12].

In the frame of this research, five oriental and six flue and air-cured type varieties were cultivated in five regions of Greece, for three years, to evaluate the yield and the physicochemical properties of the leaves. Greek varieties and production of tobacco in Greece are considered to be of the best quality, but very limited research has been focused on the yield and the quality characteristics of varieties, as well as on the impact inflicted by the area of cultivation. In traditional tobacco countries, such as Greece, farmers often may give up the crop, when the yields of tobacco are insufficient because of too much attention to quality. Thus, it is crucial to aid growers overcome the difficulties in growing and processing, in order not only to support the increase in production and quality but also to solidify the income of growers.

2. Materials and Methods

The tobacco varieties tested were separated into two groups. The first group (A oriental type) consisted of Katerini 53 (K53), Xanthi 81 (X81), Xanthi 2A (2A), Doxato, and Nigrita 34 (N34) varieties. The corresponding experimentation fields were established in the areas of Xanthi (41.06136; 24.96317), Serres (41.06149; 23.74729), Katerini (40.31725; 22.57558) Karditsa (39.34094; 21.83569) and Aitoloakarnania (38.60507; 21.48515). The second group (B flue and air-cured type) consisted of the varieties Niki, Virginia E9 (VE9), Burley 21E, Klio, NC297, and NC7L. The related experimentation fields were established in the areas of Xanthi, Karditsa, and Aitolokarnania. All the varieties tested (except NC 297 and NC7L) were registered to the Greek national catalogue of cultivated varieties from the Hellenic Agricultural Organization DIMITRA, and the breeder is the Institute of Plant Breeding and Genetic Resources. The hybrids NC 297 and NC7 were bred by the North Carolina State University and have been produced by the Gold Leaf Seed Company.

The characteristics of the Greek origin varieties, presented in

Table 1. Morphological and agronomical characteristics of Greek origin tested varieties.

Variety	Type	Plant Shape	Height 1	Number of Useful Leaves	Leaf Shape	Length/Width Ratio 2	Leaf Basis Type	Period of Harvesting (Days)
Katerini 53	sun cured–basic	conical	medium	26–30	cordate	<1.8	petiolate	46–52
Xanthi 81	sun cured–aromatic	conical	medium	26–30	elliptic	1.8–2.2	sessile with auricles	39–45
Xanthi 2A	sun cured–aromatic	conical	medium	21–25	elliptic	1.8–2.2	sessile without auricles	39–45
Doxato	sun cured–aromatic	conical	high	26–30	elliptic	1.8–2.2	sessile with auricles	46–52
Nigrita 34	sun cured–aromatic	conical	very high	21–25	elliptic	<1.8	sessile with auricles	39–45
Niki	flue cured	conical	high	26–30	elliptic	1.8–2.2	sessile with auricles	>60
Virginia E9	flue cured	conical	very high	26–30	elliptic	1.8–2.2	sessile with auricles	53–59
Klio	flue cured	conical	very high	21–25	elliptic	1.8–2.2	sessile with auricles	>60
Burley 21E	air cured	conical	very high	21–25	elliptic	1.8–2.2	sessile with auricles	39–45

¹ Medium (76–90 cm); high (91–105 cm); very high (106 cm). ² of the middle useful leaf.

, are the plant shape, the plant height (cm) from the base up to two leaves before the inflorescence, the number of useful leaves from the base up to two leaves before the inflorescence at the stage of full growth, and 50% flourished plants, the leaf shape along with the length/width ratio of the middle useful leaf, the leaf basis type with petiole or not, and the period of harvesting as days from the beginning till the end of the harvest. All these characteristics are according to the UPOV guidelines for the conduct of tests for distinctness, uniformity, and stability [13].

The experimental fields were established for three years, as a split-plot design with three replications. Each plot consisted of four rows of 10 m each with a 1 m distance between rows. The areas were chosen, in cooperation with high-expertise farmers, as the most proper for tobacco cultivation in Greece. The farmers followed the standard cultivation practices and during the growing season, four measurements were conducted, regarding leaf fresh weight per hand reaping and total leaf dry weight in Kg per ha. The harvesting was performed by hand, and the harvesting dates along with the number of hands reaping in each area are presented in

Table 2. The harvesting dates and the number of hands reaping in the tobacco cultivation areas for the years 2014, 2015, and 2016.

		Cultivation Area of Greece
	Year	Xanthi	Serres	Katerini	Karditsa	Aitoloakarnania
Transplanting	2014	2 June	30 May	15 May	25 May	3 June
	2015	8 June	25 May	18 May	11 May	18 May
	2016	10 June	28 May	20 May	29 May	25 June
1st hand	2014	10 August	21 July	28 July	10 August	18 July
	2015	2 August	13 July	14 July	27 July	10 July
	2016	25 August	27 June	27 June	30 July	20 July
2nd hand	2014	5 September	17 August	13 August	28 August	31 July
	2015	25 August	26 July	3 August	11 August	22 July
	2016	5 September	18 July	12 August	14 August	30 July
3rd hand	2014	15 October	7 September	27 August	14 September	18 August
	2015	12 September	17 August	14 August	30 August	14 August
	2016	15 September	16 August	30 August	22 August	10 August
4th hand	2014	-	2 October	12 August	-	30 August
	2015	-	9 September	3 September	10 September	28 August
	2016	-	20 September	15 September	1 September	20 August

. Mean monthly values of temperature and rainfall data during the three growing seasons, as recorded in each area (www.meteo.gr, accessed on 23 Arpil 2023), are given in Figure 1.

After the harvesting and curing, samples were used for chemical determinations. Samples were dried at 100 °C for 3 h, to constant dry weight, in a forced-air oven, and were ground in a Wiley mill using a 1 mm sieve and were measured at the laboratory of Plant Breeding and Genetic Resources Institute in Drama regarding the percentage of (a) nitrates, according to Coresta method No 36 (ISO 15152: 2002) [14], (b) nicotine percentage, with Coresta method Νο 35 (ISO 15153:2002) [15], and (c) sugars, according to Coresta method Νο 38 (ISO 15154: 2002) [16].

The primary endpoint of the statistical analysis was to assess the impact of three independent factors, (i) the cultivation year, (ii) the area/location, and (iii) the variety, on three different dependent variables, the yield, the nicotine, and the sugars. The assessment was performed, separately and independently, for (A) the oriental, and (B) the flue and air-cured type varieties. To this end, in each case, the three-factor analysis of variance (ANOVA) without replication was employed for each dependent variable, to assess these three factors as main effects, along with their pairwise interactions (the three-way interaction was not assessed). Thus, in each ANOVA performed, six hypotheses were tested in total, three related to each of the main effects, and three more related to each of the three pairwise interactions. For the hypotheses concerning the main effects, post hoc comparison was employed, when appropriate, using the Tukey HSD post hoc test. More specifically, all the pairwise comparisons were performed, and, in addition, the groups of each factor were included in different homogeneous subsets. Interaction plots were used to visualize the pairwise interactions. The Pearson correlation coefficient was computed to assess the linear relationship among the yield, and the nitrates, sugars, and nicotine within the different categories of the cultivation year, the area, and the variety. In the case of the oriental type varieties, the Pearson correlation coefficient was not computed for the area Serres due to missing values in nitrates, sugars, and nicotine. Similarly, in the flue and air-cured type varieties, it was not computed for the area Xanthi due to missing values in nitrates, sugars, and nicotine. The significance level was set to 0.05. The analysis has been performed with SPSS v22.

3. Results

3.1. A—Oriental Type Varieties

Concerning the yield, the results of the three-Factor ANOVA are shown in

Table 3. Three-factor analysis of variance for the Oriental type varieties. Yield was considered as the dependent variable and cultivation year, area, and variety as the three independent factors. Main effects, along with the pairwise interactions are assessed.

	F Statistic	p-Value
Corrected Model	4.454	<0.001
Intercept	5175.763	<0.001
Area × Variety	2.276	0.023
Cultivation year × Area	6.892	<0.001
Cultivation year × Variety	0.596	0.774
Cultivation year	11.639	<0.001
Area	6.097	0.001
Variety	10.768	<0.001

, and it was found that all three factors as main effects, were statistically significant (Cultivation year: p < 0.001, Area: p = 0.001, Variety: p < 0.001). The results of the corresponding post hoc analysis are shown in the Supplementary Materials (Supplementary Tables S4–S6). Particularly, regarding the cultivation year, statistically significant differences in the yield were observed between 2014 and both 2015 (p = 0.007), and 2016 (p < 0.001), but not between 2015 and 2016 (p = 0.322) (Supplementary Table S4A). The Tukey HSD included cultivation years 2015 and 2016 in the first homogeneous subset, while the year 2014 exhibited a much larger yield mean value and constituted by itself the second homogeneous subset (Supplementary Table S4B). Concerning the factor area, statistically significant differences were observed between Aitoloakarnania, both Serres (p = 0.002) and Xanthi (p = 0.032), and between Karditsa and Serres (p = 0.010), (Supplementary Table S5A). The three homogeneous subsets obtained in this case were not distinct from each other as in the case of the cultivation year, on the contrary, they all exhibited overlaps (Supplementary Table S5B). In the case of variety, post hoc analysis detected four statistically significant differences in the yield, which were all related to K53 (vs. DOXATO, p < 0.001; vs. N34, p < 0.001; vs. X2A, p < 0.001; vs. X81, p = 0.007) (Supplementary Table S6A). More specifically, K53 exhibited a mean value that was statistically significantly higher compared to all other varieties. This is further shown in (Supplementary Table S6B), where K53 with a yield of 1652.367 kg/ha constituted the second subset, while all other varieties were included in the first one.

Regarding the pairwise interactions of the three factors (Table 3), it was found that the area exhibited significant interaction related to yield with both variety (p = 0.023), and cultivation year (p < 0.001). The corresponding plots, where the mean yield is displayed for all combinations of the factors’ categories in each of the three pairwise interaction assessments are presented in Figure 2A–C, respectively. In Figure 2A, it is shown that the interaction between area and variety may be attributed to a large extent to the fact that variety K53 exhibited a much larger mean value in Katerini (1979.633 kg/ha), compared to the remaining four varieties in Katerini, which exhibited an overall mean value of 1309.075 kg/ha (see Supplementary Table S1 for the specific values in the remaining four varieties), while, e.g., in Aitoloakarnania, and in Xanthi, K53 exhibited a lower mean value compared to two of the remaining varieties. At the same time, the variety X81 exhibited a larger mean value in Xanthi (1631.667 kg/ha), compared to the remaining four varieties (overall mean = 1429.834 kg/ha), while in all other areas, it exhibited a lower value compared to, at least, variety K53. Similarly, X2A exhibited a larger mean value in Xanthi (1550.000 kg/ha), compared to the remaining three varieties (only excluding X81), while in the remaining areas it exhibited, in general, lower values, particularly, the lowest mean yield in Aitoloakarnania, and Katerini. In Figure 2B, it is shown that the interaction between area and cultivation year is very strong. A characteristic example is that the yield exhibited far larger mean values in 2014, compared to 2015 and 2016, in Aitoloakarnania, Karditsa, Katerini, and, Xanthi, while in Serres, it exhibited the lowest mean value (Supplementary Table S2). In Figure 2C, it is shown that the interaction between variety and cultivation year is mild, with the pattern of the mean yield across the five different varieties being similar in the three years of cultivation considered (Supplementary Table S3).

In addition, the Pearson correlation coefficient values among the yield, and the nitrates, sugars, and nicotine were computed, and the results are shown in

Table 4. The Pearson correlation coefficient for the Oriental type varieties, between the yield, and the nitrates, sugars, and nicotine, within the different categories of the cultivation year, the area, and the variety is displayed, along with the corresponding p-values.

Cultivation Year		Nitrates	Sugars	Nicotine
2014	Pearson Correlation	0.301	−0.204	−0.095
	p-value	0.197	0.388	0.689
2015	Pearson Correlation	−0.313	0.274	−0.326
	p-value	0.179	0.243	0.161
2016	Pearson Correlation	0.316	−0.501 *	0.267
	p-value	0.174	0.024	0.255
Area		Nitrates	Sugars	Nicotine
Aitoloakarnania	Pearson Correlation	0.162	−0.148	0.018
	p-value	0.565	0.599	0.948
Karditsa	Pearson Correlation	0.755 **	−0.538 *	0.225
	p-value	0.001	0.039	0.420
Katerini	Pearson Correlation	0.022	−0.050	−0.272
	p-value	0.939	0.861	0.327
Xanthi	Pearson Correlation	0.151	−0.214	0.065
	p-value	0.591	0.445	0.818
Variety		Nitrates	Sugars	Nicotine
Doxato	Pearson Correlation	0.412	−0.475	0.300
	p-value	0.183	0.119	0.343
K53	Pearson Correlation	0.124	−0.330	0.531
	p-value	0.701	0.295	0.076
N34	Pearson Correlation	0.131	0.132	−0.451
	p-value	0.686	0.682	0.141
X2A	Pearson Correlation	0.575	−0.364	−0.055
	p-value	0.051	0.244	0.865
X81	Pearson Correlation	0.508	−0.484	0.206
	p-value	0.092	0.111	0.521

** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).

. It is shown that the value of the Pearson correlation coefficient exhibited great variation between different groups of the same factor. For example, within the factor area, it was equal to 0.755 (p = 0.001) between the yield and the nitrates in Karditsa, while it was almost zero (0.022, p = 0.939) in Katerini. Also, this linear correlation between the yield and the nitrates in Karditsa was the only positive linear correlation that was found to be statistically significant. There were also two statistically significant negative linear correlations corresponding to the yield and the sugars in Karditsa (−0.538, p = 0.039), and the yield and the sugars in 2016 (−0.501, p = 0.024).

Regarding nicotine content, the results of the three-factor ANOVA are shown in

Table 5. Three-Factor analysis of variance for the Oriental type varieties. Nicotine was considered as the dependent variable, and cultivation year, area, and variety as the three independent factors. Main effects, along with the pairwise interactions are assessed.

	F Statistic	p-Value
Corrected Model	1.667	0.097
Intercept	604.408	0.000
Area × Variety	0.346	0.971
Cultivation year × Area	3.059	0.023
Cultivation year × Variety	0.719	0.673
Cultivation Year	5.364	0.012
Area	2.289	0.104
Variety	3.122	0.033

. It was found that two of the factors were statistically significant as main effects (Cultivation year: p = 0.012, and Variety: p = 0.033). The corresponding results of the post hoc analysis are shown in the Supplementary Materials (Supplementary Tables S10–S12). As far as the cultivation year is concerned, statistically significant differences in nicotine were only observed between 2015 and 2016 (p = 0.009) (Supplementary Table S10A). Concerning area, no statistically significant differences were observed (Supplementary Table S11A,B). In the case of variety, only one statistically significant difference was detected, related to K53 and X2A with p = 0.045 (Supplementary Table S12A).

Regarding pairwise interactions, the only statistically significant one was found in the case of cultivation year and area (p = 0.023) (Table 5). In the corresponding plot (Figure 3B), it is shown, in particular, that the year 2015 exhibited different patterns across the different areas, with the mean nicotine in Katerini assuming the lowest value of all 1.252%, (Supplementary Table S8). In accordance with Table 5, Figure 3A,C shows that the corresponding interactions are non-existent for area vs. both variety and cultivation year.

In

Table 6. Three-Factor analysis of variance for the Oriental type varieties. Sugars was considered as the dependent variable, and cultivation year, area, and variety as the three independent factors. Main effects, along with the pairwise interactions are assessed.

	F Statistic	p-Value
Corrected Model	6.579	0.000
Intercept	629.520	0.000
Area × Variety	1.267	0.298
Cultivation year × Area	12.937	0.000
Cultivation year × Variety	2.667	0.030
Area	14.920	0.000
Variety	4.892	0.005
Cultivation year	25.879	0.000

, the results of the three-Factor ANOVA concerning the sugar content are displayed. It was found that all factors were statistically significant as main effects (Cultivation year: p < 0.001, Area: p < 0.001, and Variety: p = 0.005). The results of the related post hoc analysis are shown in the Supplementary Materials (Supplementary Tables S16–S18). Regarding the cultivation year, statistically significant differences in the sugars were obtained between 2015, and both 2014 and 2016 (p < 0.001 in both cases) (Supplementary Table S16A). The year 2015 exhibited higher mean sugars value compared to both other years and was, thus, discerned from them, constituting a separate subset in itself (Supplementary Table S16B), while the years 2014 and 2016 were included in another subset. Concerning area, statistically significant differences were observed in the cases of Aitoloakarnania and both Karditsa (p = 0.001), and Xanthi (p < 0.001), and Katerini and Xanthi (p = 0.003) (Supplementary Table S17A). In the case of variety, two statistically significant differences were detected (K53 vs. N34, p = 0.050, and N34 vs. X2A p = 0.002, Supplementary Table S18A).

Regarding the pairwise interactions of the three factors (Table 6), it was found that the cultivation year with both area and variety exhibited significant interactions (p < 0.001, p = 0.030, respectively). In Figure 4B, it is further shown that area and cultivation year interact, particularly, the year 2016 exhibited different patterns across the different areas compared to years 2014 and 2015 (see Supplementary Table S14 for specific sugars values). Figure 4C demonstrates that the interaction found between cultivation year and variety was marginal, again mainly attributed to the year 2016, while Figure 4A shows that the corresponding interaction is not strong.

3.2. B—Flue and Air Cured Type Varieties

Concerning the yield of B-type tested varieties, the results of the three-factor ANOVA are shown in

Table 7. Three-Factor analysis of variance for the flue and air-cured type varieties. Yield was considered as the dependent variable, and cultivation year, area, and variety as the three independent factors. Main effects, along with the pairwise interactions, are assessed.

	F Statistic	p-Value
Corrected Model	8.746	<0.001
Intercept	3314.602	<0.001
Area × Variety	1.850	0.116
Cultivation year × Area	4.766	0.007
Cultivation year × Variety	0.740	0.681
Cultivation year	7.764	0.003
Area	105.024	<0.001
Variety	3.616	0.017

. It was found that all three factors as main effects were statistically significant (Cultivation year: p = 0.003, Area: p < 0.001, Variety: p = 0.017). The results of the post hoc analysis regarding the main effects are shown in the Supplementary Materials (Supplementary Tables S22–S24). More specifically, in the case of cultivation year, the year 2014 exhibited the highest mean yield value 3343.889 kg/ha—(Supplementary Table S22B), and statistically significant differences compared to both other years (vs. 2015: p = 0.006, vs. 2016: p = 0.009, Supplementary Table S22A). Consequently, the Tukey HSD included cultivation years 2015 and 2016 in the first homogeneous subset, and the year 2014 in the second homogeneous subset (Supplementary Table S22B). Concerning the factor area, two statistically significant differences were observed (Aitoloakarnania vs. Karditsa: p < 0.001, and Xanthi vs. Karditsa: p < 0.001, Supplementary Table S23A). More specifically, Karditsa exhibited by far the lowest mean yield value 1967.788 kg/ha (Supplementary Table S23B). Thus, Karditsa constituted a separate homogeneous subset, based on Tukey HSD, while Aitoloakarnania and Xanthi constituted the second one (Supplementary Table S23B). As for varieties, the only statistically significant difference was observed when comparing the variety Burley 21E with NC297 (p = 0.007) (Supplementary Table S24A). Thus, the corresponding Tukey homogeneous subsets exhibited great overlap (Supplementary Table S24B).

By evaluating the pairwise interactions (Table 7), it was found that the only interaction that was found to be statistically significant was between the cultivation year and area (p = 0.007). The corresponding plots, where the mean yield is displayed for all combinations of the factors’ categories are presented in Figure 5A–C, respectively. In Figure 5A, it is shown that the interaction between area and variety is mild. The different varieties follow the same pattern across the three areas with the exception of the yield mean value that is observed in Aitoloakarnania for the variety Burley 21E, which appears to be much lower than in the remaining varieties (see also Supplementary Table S19). In Figure 5B, it is visually shown how area and cultivation year interact, in particular, the mean yield value in Aitoloakarnania during 2016 (3163.317 kg/ha, Supplementary Table S20) was much lower than expected in case no interaction was present.

The results of the Pearson correlation coefficient values among the yield of B type varieties, and the nitrates, sugars, and nicotine were computed and are displayed in

Table 8. The Pearson correlation coefficient for the flue and air cured type varieties, between the yield, and the nitrates, sugars, and nicotine, within the different categories of the cultivation year, the area, and the variety is displayed, along with the corresponding p-values.

Cultivation Year		Nitrates	Sugars	Nicotine
2014	Pearson Correlation	0.137	0.210	0.710 **
	p-value	0.672	0.513	0.010
2015	Pearson Correlation	0.599 *	−0.148	0.040
	p-value	0.039	0.645	0.901
2016	Pearson Correlation	0.133	0.441	−0.608 *
	p-value	0.680	0.152	0.036
Area		Nitrates	Sugars	Nicotine
Aitoloakarnania	Pearson Correlation	−0.622 **	0.376	0.301
	p-value	0.006	0.124	0.224
Karditsa	Pearson Correlation	−0.057	−0.139	−0.022
	p-value	0.824	0.582	0.930
Variety		Nitrates	Sugars	Nicotine
Burley 21E	Pearson Correlation	0.363	−0.841 *	0.139
	p-value	0.479	0.036	0.793
KLIO	Pearson Correlation	0.676	0.373	−0.009
	p-value	0.141	0.467	0.986
NC297	Pearson Correlation	0.784	−0.355	0.039
	p-value	0.065	0.490	0.942
NC7LC	Pearson Correlation	−0.379	−0.623	−0.049
	p-value	0.458	0.186	0.927
NIKI	Pearson Correlation	0.340	0.756	−0.770
	p-value	0.510	0.082	0.073
VE9	Pearson Correlation	0.859 *	−0.323	−0.110
	p-value	0.029	0.532	0.836

** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).

. It is shown that the value of the Pearson correlation coefficient exhibited great variation. For example, within the factor area, it was equal to −0.622 (p = 0.006) between the yield and the nitrates in Aitoloakarnania, while it was almost zero (−0.057, p = 0.824) in Karditsa. Within the cultivation year, it was found that the yield and the nicotine were negatively linearly correlated in 2016 (−0.608, p = 0.036), while they were positively linearly correlated in 2014 (0.710, p = 0.010). The strongest positive linear correlation was obtained between the yield and the nitrates within the variety VE9 (0.859, p = 0.029). On the other hand, the strongest negative linear correlation was obtained between the yield and the sugars within the variety Burley 21E (−0.841, p = 0.036).

In

Table 9. Three-Factor analysis of variance for the flue and air-cured type varieties. Nicotine was considered as the dependent variable, and cultivation year, area, and variety as the three independent factors. Main effects, along with the pairwise interactions are assessed.

	F Statistic	p-Value
Corrected Model	2.375	0.077
Intercept	947.013	0.000
Area × Variety	1.223	0.366
Cultivation year × Area	12.343	0.002
Cultivation year × Variety	0.883	0.576
Cultivation year	5.426	0.025
Area	5.359	0.043
Variety	0.708	0.631

, the results of the three-Factor ANOVA concerning the % nicotine content of B type varieties are presented. It was found that two of the factors were statistically significant as main effects (Cultivation year: p = 0.025, and Area: p = 0.043). The corresponding results of the post hoc analysis are shown in the Supplementary Materials (Supplementary Tables S28 and S29). In the case of cultivation year, statistically significant differences in nicotine were only observed between 2014 and 2016 (p = 0.022) (Supplementary Table S28A). In the case of area, only two areas were assessed (Aitoloakarnania and Karditsa), thus post hoc analysis was not performed. Concerning variety, no statistically significant results were obtained (Supplementary Table S29).

Regarding pairwise interactions, it was found that only cultivation year and area exhibited significant interaction (p = 0.002, Table 9). This is clearly shown in the corresponding Figure 6B, with the year 2016 exhibiting different value pattern across areas compared to 2014 and 2015. Particularly, during 2016 in Karditsa, nicotine assumed its highest mean value 1.892% (Supplementary Table S26).

The results of the ANOVA concerning the sugars content (%) of B-type varieties are shown in

Table 10. Three-factor analysis of variance for the flue and air cured type varieties. Sugars was considered as the dependent variable, and cultivation year, area, and variety as the three independent factors. Main effects, along with the pairwise interactions are assessed.

	F Statistic	p-Value
Corrected Model	14.673	<0.001
Intercept	1294.624	<0.001
Area × Variety	6.109	0.008
Cultivation Year × Area	8.955	0.006
Cultivation Year × Variety	2.147	0.122
Cultivation Year	6.682	0.014
Area	1.135	0.312
Variety	56.478	<0.001

. It was found that two factors were statistically significant as main effects (Cultivation year: p = 0.014, and Variety: p < 0.001). The results of the post hoc analysis regarding the main effects are shown in the Supplementary Materials (Supplementary Tables S33 and S34). Regarding the cultivation year, statistically significant differences in the sugars were observed between 2015, and both 2014 and 2016 (p = 0.030, and p = 0.021, respectively) (Supplementary Table S33A). The year 2015 exhibited larger mean sugars value (10.445%) compared to both other years and was, thus, discerned from them, constituting a separate Tukey HSD subset in itself (Supplementary Table S33B). In the case of area, since only two areas were assessed, post hoc analysis was not performed. Concerning variety, statistically significant differences were observed in ten cases, (see Supplementary Table S34A). As far as the pairwise interactions of the three factors are concerned (Table 10), it was found that the area and both variety, and cultivation year exhibited significant results (p = 0.008, and p = 0.006, respectively), as it is also shown in Figure 7A,B.

4. Discussion

Tobacco is a plant grown for its leaves and is an important crop for farmers in Greece through the years [17]. The characteristics of tobacco genotypes are the consequence of the common effect of the genotype, environment, and their interaction [18]. The results of this study indicate the differences between the varieties regarding their yield and quality in different regions of Greece and it serves to identify the components of tobacco thought to be influential in determining the quality parameters of tobacco leaves.

Dried leaf yield has been the characteristic that interests not only plant breeders but also tobacco producers [19]. For A type varieties (Supplementary Table S1), K53 variety in Katerini area exhibited its maximum yield (1979.63 kg/ha), followed by Serres (1815.77 kg/ha) and Karditsa (1672.10 kg/ha), while in Aitoloakarnania the higher yield was observed by the variety Doxato (1372.77 kg/ha). In the area of Xanthi, the X81 variety had the optimum yield (1631.67 kg/ha) followed by X2A (1550.00 kg/ha). The genetic potential of each tobacco variety theoretically exceeds some limits in the appearance of a particular trait [20]. Thus, the higher yields of the varieties K53 and X81 were observed in Katerini, and in Xanthi areas, respectively, which was expected and is mainly explained because of the variety adaptation in the above regions [7,21].

Regarding the B type varieties (Supplementary Table S19), Aitoloakarnania area exhibited high yields during all the years of experimentation, with the higher values obtained in the case of NC297 (4058.00 kg/ha), followed by Klio (4044.67 kg/ha), and NC7LC (3799.00 kg/ha). Only the yield of variety Burley 21E (2673.63 kg/ha), appears to be much lower than the remaining varieties in the above region. In Xanthi area, the most productive variety was NIKI (3626.67 kg/ha) and in Karditsa the NC297 (2390.67 kg/ha). The results showed that in areas located in the central and south part of Greece, such as Karditsa and Aitoloakarnania, B type varieties demonstrated greater adaptability.

During the current experimentation, both A and B type varieties reached their optimum yield in 2014 cultivation year, with statistically significant differences with the other two years. The data are in accordance with the results of [13], where the interaction of the year with the cultivar was also significant for total yield and nicotine content.

The aroma and basic quality characteristics of tobacco largely depend on the climate of the region which affect the performance of plant growth, productivity, and chemical composition of tobacco plants [22]. Through the experimentation years of this study, the Pearson correlation coefficient values for oriental type varieties shown that there is no statistically significant correlation in the experimental fields between yield and content of nitrates and sugars (except for Karditsa), while for B type of varieties the strongest positive linear correlation was obtained between the yield and the nitrates in variety VE9, and the strongest negative linear correlation was obtained between the yield and the sugars within the variety Burley 21E. According to [23], time selection for leaf harvest, which is defined by mature period, is important for yield and quality of tobacco leaves, despite the great variability of soil and climatic conditions.

Oriental tobaccos are famous for their quality characteristics and the high aroma of their small leaves with low sugar and nicotine contents but the relationship between the many components of tobacco and tobacco quality is complex [12]. Our experimentation showed that the nicotine level of the X2A variety was the highest during all years, and in all areas of experimentation, and ranged from 2.19 to 3.05% (Supplementary Table S9). The only exception was the area of Xanthi where the variety Doxato exhibited the highest value of nicotine 2.83% (Supplementary Table S7). The variety X81 followed with nicotine levels ranging from 1.77 to 2.60% (Supplementary Table S9), which is also in agreement with [20], where X81 had statistically significantly higher nicotine when compared to the other tested varieties (K53 and Myrodata Agriniou). In Turkey, nicotine contents of dried oriental tobacco leaves have been reported to vary from 1.5 to 3.5% [24], and from 2.1 to 3.3% [25], while [26] found nicotine content values of oriental tobacco to lie within 0.96–2.06% in other regions of Turkey, and 0.47% in Samsun region. The chemical composition of tobacco leaves was significantly different among locations where ecological conditions such as light, humidity, precipitation, altitude, and temperature were different [27], and previous results have shown that the chemical characteristics of tobacco are affected by soil and climate, except for the genetic potential [26]. The content of nicotine and soluble carbohydrates is positively correlated with temperature and negatively with rainfall [28]. Oriental tobaccos have a different flavor, and lower nicotine levels compared to commercial flue-cured and air-cured tobaccos, while the most important agronomic decision affecting nicotine levels in flue-cured tobacco leaves is the variety choice [11]. As reported in [29], the oriental tobaccos have better quality characteristics compared to Virginia and Burley tobaccos. The results of our experiments showed that the mean nicotine content % of B-type varieties is around 1.3% with a range of 0.66–2.12% (Supplementary Table S27), lower than A type. Τhe differences can be explained by the fact that the tobacco genotypes and cultivars showed differences depending on their adaptation, genetic properties, and survival in different places. In addition, the soil contents as organic matter can affect the nicotine of tobacco [30].

Tobacco leaves’ sugar content is an important measure of chemical quality in tobacco and is directly related to the taste and flavor of tobacco. In general, tobacco varieties with high sugar contents are considered to be, of better quality [31]. Leaf glucose content in oriental tobacco leaves was reported to be 2.0% [32], 2.98% [27], and 4.2% [28]. The sugar content values for oriental-type varieties, in the present study, were higher than those reported in the previous referred studies and ranged from 2.71 to 9.78% (Supplementary Table S15). The only exceptions were the Karditsa area in 2014 with a sugar content of 1.21%, and the Xanthi area in 2016, which exhibited an extremely low value of 0.51% (Supplementary Table S14). The year 2015 demonstrated the highest sugar content of all other years (Supplementary Table S14), while Aitoloakarnania and Katerini regions resulted in higher values of sugars (Supplementary Table S13). For the B-type varieties, VE9 had the higher content of all tested varieties (15.06% sugars in Karditsa and 14.51% in Aitoloakarnania), while Burley 21E and NC7LC had the lowest (Supplementary Table S30) sugar content.

Previous studies in Greek tobacco varieties, which took place in the Tobacco Institute of Drama [33] showed for Basma type variety nicotine content of 2–3% and sugars 10–12%, while the K53 variety had nicotine of 1.5–2.0% and sugars 18%. For the Virginia variety nicotine content was 2.1% and sugars 17%, while for the Burley variety 2% and 0%. Although the above experimentation has been materialized many years ago, these results are in accordance with the current results, implying that nicotine and sugar content is mainly related to the genotype of tobacco, and secondly to agronomical practices [34]. Moreover, the present study showed that the Greek varieties registered to the Greek National catalogue were stabilized and their quality parameters are not only unchanged but, in addition, they fulfill the task for new varieties and hybrids, mentioned by [28], for reducing sugar content of 8–12% and nicotine content of 2.2–2.7%. Further studies should be conducted on the evaluation of the effects of several agronomic practices on growth, yield, and quality, especially for the less studied varieties of tobacco or new hybrids that gain marketplace.

5. Conclusions

Variety selection is the most important aspect for achieving the desired yield and quality of tobacco and the results of this research confirmed that variety adaptation in particular regions resulted in higher yields. Referring to reducing sugar and nicotine contents of cured leaves, the major chemical quality properties of tobacco, the A-type varieties exhibited higher levels of nicotine compared to B type and reducing sugars content was found to be at the optimum desirable level. Τhis study would be helpful, for tobacco producers and users, to understand the chemical contents in Greek tobacco varieties, which are cultivated and known worldwide. This study would be helpful, for tobacco producers and users, to understand the chemical contents in Greek tobacco varieties, which are cultivated in specific areas and are known worldwide. This need is intensified by the fact that recent and proposed changes in the tobacco industry may require more precise management of tobacco quality mainly through the cultivation of varieties well adapted to specific environments.