Salt tolerance during the seedling production stage of Catharanthus roseus, Tagetes patula and Celosia argentea1 Tolerância à salinidade na produção de mudas de Catharanthus roseus, Tagetes patula e Celosia argentea

The guarantee of water supply for irrigated agriculture in the semi-arid region must necessarily involve the use of lower quality water, such as brackish water. The objective of the present work was to evaluate the tolerance to salinity of the ornamental species Catharanthus roseus, Tagetes patula and Celosia argentea, using different methods. The experiment was conducted in a randomized block, arranged in a 10 x 3 factorial scheme, corresponding to 10 saline concentrations of irrigation water (ECw 0.5; 1.0; 1.5; 2.0; 2.5; 3.0; 3.5; 4.0; 5.0 and 6.0 dS m-1) and 3 ornamental species. Four salinity tolerance assessment methods were tested, using relative values or percentages of reduction in quantitative and qualitative analyses. The different methods show the highest sensitivity to salinity of C. roseus, in the seedling production stage, compared to T. patula and C. argentea species. The methods of threshold salinity and ORN index led to similar results in terms of classification of salt tolerance, with C. roseus classified as sensitive and T. patula and C. argentea as moderately sensitive. The method of Fageria (1985) allowed good separation of the species, with tolerance limits of 1.5, 2.5 and 3.5 dS m-1, respectively for C. roseus, T. patula and C. argentea. It is obvious from the comparison with literature data that the seedling production stage is more sensitive to salt stress, and it is necessary to carry out new studies aimed at attenuating the effects of stress at this stage, through the use of management techniques.


INTRODUCTION
One of the great challenges of contemporary agriculture is directly related to the water issue because, given the effect of climate change and the lack of more effective policies for recycling water, it tends to become increasingly limited both qualitatively and quantitatively. In this context, it is necessary to advance within the possibilities of using lower quality water in agricultural production. In this perspective, special attention should be paid to brackish water, which normally does not require any type of chemical treatment and its use in agriculture depends on the adoption of a set of management techniques, especially associated with the establishment of reference indices of salinity tolerance for crops (GARCIA-CAPARRÓS;LAO, 2018;LACERDA et al., 2016).
The methods used to classify plant tolerance to salinity presuppose the existence of enormous intraspecific and interspecific genetic variability, which may result in species or varieties with low, intermediate or high capacity to withstand excess salts in the cultivation environment (DIAS et al., 2016;SOARES FILHO et al., 2016). This level of tolerance also depends on the development stage of the plant and on other factors, such as type of salt, irrigation method and frequency, and climatic conditions (MEDEIROS et al., 2016).
Among the methods adopted to classify plant responses to salinity, those that are mainly based on growth and important agronomic characteristics, such as grain and fruit production, stand out (AYERS; WESTCOT, 1999), considering values of salinity threshold for relative yields (MAAS; HOFFMAN, 1977) or percentages of reduction in growth or yield (FAGERIA, 1985;MIYAMOTO et al., 2004).
The application of these methods is widely recognized, but little is known in terms of comparison between them, especially in studies focused on tolerance to salinity of ornamental plants (OLIVEIRA et al., 2018). Although plants are commonly grouped into divisions of salinity tolerance based on biometric and production parameters, for ornamental plants a separation based also on visual quality may be the most appropriate (CASSANITI; ROMANO;FLOWERS, 2012;NEVES et al., 2018;OLIVEIRA et al., 2018).
Salinity tolerance studies with ornamental plants have been carried out mainly after they have been transferred to the field, in order to evaluate the capacity of establishment of seedlings under irrigation with saline water (ALVARÉZ; SÁNCHEZ-BLANCO, 2015;CARILLO et al., 2019;LEONARDI;FLOWERS, 2009;ESCALONA et al., 2014;FARIERI et al., 2016;GARCIA-CAPARRÓS et al., 2016;MIYAMOTO et al., 2004;NIU;STARMAN;BYRNE, 2013;NEVES et al., 2018;OLIVEIRA et al., 2017;OLIVEIRA et al., 2018;VALDEZ-AGUILAR;GRIEVE;POSS, 2009;VEATCH-BLOHM;ROCHE;SWEENEY, 2019). On the other hand, there are few studies aiming to evaluate the tolerance of these plants in the seedling production stage, which is possibly more sensitive to excess salts. These studies can generate information of great relevance to support the use of saline waters by small companies and farmers working in the ornamental sector, indicating the levels of water salinity that can be used in the production of seedlings.
In this context, the present work aimed to evaluate the salinity tolerance of three ornamental species (Catharanthus roseus -'Boa noite', Tagetes patula -'Cravo amarelo' and Celosia argentea -'Crista de galo') in the seedling production stage, using qualitative and quantitative analyses.

Research site location
The experiment was conducted in a greenhouse, located in the experimental area of the Agrometeorological Station of the Federal University of Ceará in Fortaleza (3º45' S; 38º33' W), Ceará, Brazil, from July to September 2018. Data of temperature, relative humidity and sunlight level were collected every hour using a datalogger (Onset -Hobo). The mean air temperature ranged from 28.1 to 31.2 ºC, while the relative humidity ranged from 57.2 to 65.7%, and the mean values of daily light ranged from 13,973.9 to 22,729.8 Lux. The photoperiod was approximately 12 h during the experimental period.

Experimental design and preparation of solutions
The experimental design was in randomized blocks, with four replicates, in a 10 x 3 factorial scheme, corresponding to 10 electrical conductivities of irrigation water (ECw 0.5; 1.0; 1.5; 2.0; 2.5; 3.0; 3.5; 4.0; 5.0 and 6.0 dS m -1 ) and 3 species of ornamental plants, Celosia argentea, Tagetes patula and Catharanthus roseus ( Figure  1), totaling 120 experimental units, each formed by three pots containing one plant each.

Sowing, application of treatments and cultural practices
Sowing was performed directly in polyethylene pots, with capacity of 700 mL, placing on average five seeds per pot. These were filled with substrate composed of a mixture of crushed and sifted carnauba bagana (leaf fiber byproduct from wax production), earthworm humus and arisco (sandy material with light texture normally used in constructions in Northeast Brazil), in the ratio of 2:1:1. The size of the pots and the substrate used were defined based on information from ornamental producers of the region. The substrate was subjected to irrigation, which brought it to the saturation condition, followed by drainage of excess water to reach its field capacity.
Treatment application started seven days after sowing (DAS), when plants had emerged. Thinning, keeping one plant per pot, was performed at 14 DAS, and then fertilization was carried out with N-P-K in the 10-10-10 formulation, applying 1.0 g per pot. Irrigation management was performed by water balance, according to equation 01, maintaining one pot as a drainage lysimeter for each species and each salinity level. A leaching fraction of 0.15 was added in each irrigation event in order to avoid excessive accumulation of salts in the root zone (AYERS; WESTCOT, 1999). At the end of the study, the total volumes of water applied for each salt level (36 pots), considering 28 irrigation events, were: 170. 28, 171.00, 164.10, 164.88, 160.20, 156.48, 149.40, 146.76, 127.44, 120.78 L, respectively for 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 5.0 and 6.0 dS m -1 . Where: TIR -Total irrigation required in mL; VA -Volume applied to the lysimeter in mL; VD -Volume drained in mL; LF -Leaching fraction.

Evaluations
At 47 days after the imposition of treatments (DAIT), sensory and biometric analyses (plant height, stem diameter, number of flowers and shoot dry mass) were performed.

Sensory analysis
Initially, 30 plants were randomly selected, one plant per salinity level and per species, and subjected to sensory analysis according to the methodology described by Ureña, D'árrigo and Girón (1999), adapted for qualitative evaluation of the effects of salinity (NEVES et al., 2018). The hedonic scale with nine numerical points with the following limits was used: one (Disliked extremely) and nine (Liked extremely). The analysis was performed by 100 randomly chosen judges, consisting of students, employees and professors from the Federal University of Ceará -UFC, and 83, 12 and 5% of the public were aged 18-35, 36-55 and 56-70, respectively, distributed in 68% male and 32% female. The plants selected for this evaluation had their treatment labels replaced by randomized numbers with samples separated by species, but at random positions. The scores attributed to the plants by the judges were converted into weighted averages and constituted the variable general appearance of the plants (GA).

Biometric analysis
Biometric analyses were performed, first by measuring plant height (PH, cm), stem diameter (SD, mm) and counting the number of flowers (Nflower, unit plant -1 ). Subsequently, the plants were collected, divided into flowers, leaves and branches, separately packed in paper bags and dried in an oven with forced air circulation at 65 ºC until they reached constant weight, to obtain the dry biomass of all parts, which together constituted the shoot dry mass (SDM, g plant -1 ).

Methods of classification of salinity tolerance
All quantitative and qualitative data were previously subjected to analysis of variance in order to verify single effects and/or interactions between factors (salinity and species). In order to perform the classification of the salinity tolerance, all the results were expressed in relative values, considering the control (0.5 dS m -1 ) as a reference (100% or 0% reduction).
Based on the biometric variables and the general appearance of the plants, the salinity tolerance classification of the three species was performed, using four methods.

Miyamoto et al., (2004)
The method defined by Miyamoto et al. (2004), considers a 25% reduction in the different variables evaluated. As it is irrigation water, a ratio of approximately 1.5 was used to convert ECw into ECse, considering a leaching fraction of 0.15 and substrate with medium texture (AYERS; WESTCOT, 1999). According to these criteria, the plants were classified into the following categories: sensitive (ECw from 0.0 to 2.0 dS m -1 ); moderately sensitive (ECw from 2.0 to 4.0 dS m -1 ); moderately tolerant (ECw from 4.0 to 6.0 dS m -1 ); and tolerant (ECw ˃ 6.0 dS m -1 ).

Fageria (1985)
In the method described by Fageria (1985), the percentages of reduction in the values of the studied variables were calculated using the treatment of lowest salinity (0.5 dS m -1 ) as a reference for the others. According to this criterion, the plants were classified as: tolerant (reductions from 0 to 20%), moderately tolerant (20.1 to 40%), moderately sensitive (40.1 to 60%), and sensitive (reduction of more than 60%).

Oliveira et al., (2018)
Finally, aiming to test a specific salinity tolerance index for ornamental plants, called ornamental index (ORN index), the cumulative reductions in shoot dry mass and the general appearance of the plants and in shoot dry mass and number of flowers were considered, adopting a reduction of 25%. The lowest level of salinity (0.5 dS m -1 ) was used as a reference for the other treatments, in order to express results in relative terms. According to this criterion, the plants were classified into the following categories: sensitive (ECw from 0.0 to 2.0 dS m -1 ); moderately sensitive (ECw from 2.0 to 4.0 dS m -1 ); moderately tolerant (ECw from 4.0 to 6.0 dS m -1 ); and tolerant (ECw ˃ 6.0 dS m -1 ) (OLIVEIRA et al., 2018). Table 1 shows the summary of the statistical significance of the analysis of variance for the variables shoot dry mass (SDM), stem diameter (SD), plant height (PH), number of flowers (Nflower) and general appearance of the plants (GA). All variables were influenced by the factors species and salinity individually, but SDM, Nflower ns, not significant; **P<0.01; *P<0.05. Source: created by the author and GA were also affected by the interaction between treatments (Sp x Sal), which did not occur for SD and PH. In addition, there was significant effect of the block for SD, PH and GA, evidencing the pertinence of the control of environmental variability promoted by the statistical design in randomized blocks for most variables.

RESULT AND DISCUSSION
The relative data of shoot dry mass, general appearance of plants and number of flowers show that there is similarity in the quantitative and qualitative responses of the three species studied, and that such similarity was dependent on the salinity level employed (Figure 2). For C. roseus (Figure 2A), there were linear reductions in SDM, Nflower and GA on the order of 14.05, 13.44 and 6.3% per unit increment in ECw (dS m -1 ), respectively, showing less intense effect of salt stress on the general appearance of the plants. These results differ from those obtained by other authors (NEVES et al., 2018;OLIVEIRA et al., 2018), who observed an increase in flower production and sensory evaluation up to about 2.5 dS m -1 . However, it should be noted that these studies were carried out in the stage of seedling establishment in the field and not in the production of seedlings, which suggests that the initial stage of development of C. roseus is more sensitive to salinity.
The species T. patula ( Figure 2B) also showed linear reductions in SDM and GA, equal to 13.69 and 11.96% per unit increment in ECw (dS m -1 ), respectively, whereas for Nflower, there was a quadratic response, with a maximum value observed at 1.2 dS m -1 , 20.86% above the value referring to the treatment of lowest salinity, and the values were below 100% only from the ECw of 3.0 dS m -1 . Such stability in flower production is similar to that observed in other studies with ornamental plants, which showed slightly higher values under saline conditions, which is a positive factor in the qualitative evaluation of these plants (CAI et al., 2014;NEVES et al., 2018).
For the species C. argentea ( Figure 2C), the three variables decreased linearly as a function of the increase in salinity, with higher intensity for the number of flowers and shoot dry mass, which decreased by 12.29 and 11.77% per unit increment in ECw (dS m -1 ), respectively, whereas for GA, this reduction was 8.39%. Carter et al. (2005), found a reduction in flower production in C. argentea only when irrigation water salinity was higher than 8.0 dS m -1 under southern California summer conditions. Table 2 presents the classification of salinity tolerance according to Mass and Hoffman (1977), Miyamoto et al. (2004), and Oliveira et al. (2018). In general, considering all the classification methods evaluated, it can be verified that the species C. roseus was the one which showed higher sensitivity to salt stress in the seedling production stage compared to T. patula and C. argentea. However, it can be observed that there was a discrepancy in the classification between the methods employed and between the variables, and the method of Miyamoto et al. (2004), had the most discrepant results.
According to the adapted method of Miyamoto et al. (2004), and considering the parameter SDM, the three ornamental plant species were classified as moderately sensitive to salinity, but this classification did  HOFFMAN, 1977) and in the combination of quantitative and qualitative responses (OLIVEIRA et al., 2018), the species C. roseus was classified as sensitive, while T. patula and C. argentea were predominantly classified as moderately sensitive. It is important to note that, for the method of Maas and Hoffman (1997), the variable number of flowers was the only one that showed discrepancy in the classification compared to the other variables, when the species T. patula and C. argentea were considered moderately tolerant, with salinity threshold higher than 2.0 dS m -1 .
Differently from the result obtained in the present study, Oliveira et al. (2018), compared methods to evaluate the tolerance of ornamental plants to salinity and found moderate sensitivity for C. roseus when considering the ORN index. Likewise, Friedman et al. (2007), applying secondary treatment effluents with EC of 2.3 dS m -1 for the cultivation of Celosia argentea as a cut flower, found that there was no negative interference of salinity in plant growth or flower production, suggesting that this species has moderate tolerance to salinity. However, for the species T. patula, the results obtained here were similar to those reported by Sun et al. (2018), who evaluated the Table 2 -Classification of salinity tolerance for Catharanthus roseus, Tagetes patula and Celosia argentea based on shoot dry mass (SDM), stem diameter (SD), plant height (PH), number of flowers (Nflower) and general appearance (GA) by the adapted methods of Mass and Hoffman (1977), Miyamoto et al. (2004) and Oliveira et al. (2018) S -Sensitive; MS -Moderately sensitive; MT -Moderately tolerant; T -Tolerant; * -Salinity threshold. Source: created by the author responses of cultivars of this species under conditions of irrigation with saline water and found moderate sensitivity to salinity for all cultivars studied.
The method proposed by Fageria (1985), whose results are found in Table 3, differs from the previous ones, because it enables the evaluation of the tolerance for each level of salinity tested, being easy to apply. For C. roseus, considering SDM, it was found that the species was tolerant to the effects of salinity up to the level of 3.0 dS m -1 . However, considering the relevance of the other variables, especially qualitative variables, moderate losses in the production of its seedlings already occur under salinity of 1.5 dS m -1 , with greater impacts on flower production. Neves et al. (2018), also using the adapted method of Fageria (1985), classified C. roseus as tolerant to salinity levels of up to 2.5 and 7.5 dS m -1 , considering the production of shoot biomass and flowers, respectively. However, the authors worked with seedlings already produced, that is, the study was carried out in the stage of seedling establishment under field conditions and not in the seedling production stage. This result reinforces the previous observations that the stage of production of seedlings of this species is more sensitive to salt stress, and it is necessary to establish management practices in order to ensure the production of more vigorous and better quality seedlings of this species and other ornamental  The species T. patula showed salinity tolerance limits at the level of 2.5 dS m -1 , with little divergence between qualitative and quantitative variables. However, these results differ from those obtained by Valdez-Aguilar, Grieve and Poss (2009), who conducted research to evaluate the influence of salinity and alkalinity of irrigation water on the development of three varieties of Tagetes sp. and obtained as a result the possibility of producing this species, without significant losses in plant appearance and quality, using irrigation water of up to 8.0 dS m -1 . These authors, however, started the application of saline treatments at 37 days after sowing, that is, the evaluations did not occur in the seedling production stage, which seems to be the most critical from the point of view of sensitivity to salt stress.
Finally, the species C. argentea showed salinity tolerance limits of 3.5 dS m -1 . Results higher than those found in our study were verified by Carter et al. (2005), who studied the production and absorption of ions by two cultivars of C. argentea irrigated with saline wastewater and claim that it is possible to produce them commercially with EC between 10 and 12 dS m -1 . The divergences with the results of the present study can be justified in part by the different genetic materials used. In addition, the cited authors worked with sand tanks and, during the first 20 days, irrigation was performed with a complete nutrient solution, with electrical conductivity of 2.5 dS m -1 , which was treatment considered as control. Such condition may have guaranteed the production of much more vigorous plants before being subjected to treatments of higher salinity, not adequately representing the condition of seedling production in pots under the conditions of Brazilian producers.
Adaptability to salt stress may be different between and within species belonging to the same genus or even between cultivars of the same species (CASSANITI; ROMANO;FLOWERS, 2012;DIAS et al., 2016). Although morphophysiological and biochemical processes result in adaptive responses to salt stress (ACOSTA-MOTOS et al., 2015), these are influenced by plant development stage, climatic conditions such as relative humidity and temperature, irrigation frequency, leaching fraction and soil water retention characteristics (MEDEIROS et al., 2016). Then, the same species or cultivars of this species, subjected to irrigation with similar saline water, in different regions and under different crop conditions, may exhibit divergent levels of salt tolerance.

CONCLUSIONS
1. The different methods show the greater sensitivity to salt stress of the species C. roseus, in the seedling production stage, compared to T. patula and C. argentea. The methods of salinity threshold and ORN index showed similar results in terms of salinity tolerance classification, with C. roseus classified as sensitive and the species T. patula and C. argentea classified as moderately sensitive. In average terms, the method of Fageria (1985) allowed good separation of species, with tolerance limits of 1.5, 2.5 and 3.5 dS m -1 , respectively for C. roseus, T. patula and C. argentea, being an information of easy understanding and application by the producers of seedlings; 2. Based on the analysis of the results, it is also possible to conclude that it is essential to consider the qualitative aspect at the time of classification with respect to the salinity tolerance of ornamental species, given the relevance of this aspect to the consumer; 3. The three ornamental species studied have potential for production with saline waters, paying attention to the limits observed in this study, and they may be an option for using these waters in agricultural production in the semi-arid region. However, it is obvious by the comparison with data from the literature that the seedling production stage is more sensitive to salt stress, and it is necessary to conduct further studies aimed at mitigating the effects of salinity in this stage, by employing management techniques.