Controlled deterioration test and use of the Seed Vigor Imaging System

Vigor tests are important tools for verifying the physiological potential of seed lots; however, various aspects can interfere in the consistency of the test results. The search for less subjective procedures has drawn the interest of researchers. The aim of this study was to adapt the methodology of the controlled deterioration test and verify the efficiency of the Seed Vigor Imaging System (SVIS®) to evaluate the vigor of crambe seeds. Initially, seed moisture content was determined and germination and vigor were evaluated in five seed lots. For the controlled deterioration test, the seed moisture content was adjusted to 18%, 20%, and 22%, and the results were interpreted at four and five days after sowing. The controlled deterioration test should be performed with initial adjustment of seed moisture content to 18%, and the germination test should be evaluated on the fifth day after sowing. SVIS® provided information on the vigor index and the length and uniformity of seedling development. Computerized image analysis is effective in evaluating the vigor of crambe seeds, and the uniformity of seedling development parameter offers the most consistent data.


Introduction
Crambe [Crambe abyssinica (Hochst.Ex R. E. Fr.)] is a Brassicaceae traditionally produced in the United States and Europe.The main use of the species worldwide is for production of biomass for animal feed (it has 30-32% crude protein) and for extraction of erucic acid, a monounsaturated omega-9 fatty acid that can be used in a way similar to mineral oils, but with the advantage of being easily biodegradable (Desai, 2004).According to Paulose et al. (2010), the crop stands out for high biomass production, short cycle, and hardiness.In Brazil, the crambe crop was introduced at the beginning of the 1990s by the Fundação MS, mainly with a view toward biodiesel production because the seeds have from 26% to 38% oil content (Jasper et al., 2010;Pitol et al., 2010).In spite of the agricultural potential of this species, studies directed to clarification of production characteristics are still rare.
Within this context, one of the main requisites for achieving the maximum productive potential of a crop is obtaining an adequate plant population in the field, which requires the use of high seed quality (Leão et al., 2016).From the physiological point of view, the development of standardized, fast, objective, and reliable methods is fundamental for creating seed quality control programs in seed companies (Marcos-Filho, 2010).
Traditionally, germination and vigor tests are performed to characterize seed physiological potential.The controlled deterioration test is a worthwhile alternative for avoiding the effects of accentuated differences in the water absorption rate of seed lots of the same species, especially when working with small seeds (Dutra and Medeiros-Filho, 2008).In this test, the seed moisture content is standardized to pre-established values before the period of exposure to stress.That way, seeds are artificially aged under high temperatures; however, with similar moisture content (AOSA, 2002;Lima and Marcos-Filho, 2011;Torres et al., 2013).
Studies that evaluate the best procedure for adjusting the initial seed moisture content have been performed in an attempt to refine standardization of the controlled deterioration test (Torres and Paiva, 2009;Silva and Vieira, 2010;Torres et al., 2012).Nevertheless, species or even cultivars may exhibit different responses when subjected to the procedures recommended and used for other species.
At the same time, the development of innovative alternatives has arisen with the aim of obtaining important information that is less subjective and more reliable concerning the seed physiological potential.An important alternative is the Seed Vigor Imaging System, SVIS ® , software that allows vigor index and length and uniformity of seedling development to be obtained (Kikuti and Marcos-Filho, 2012).
The methodology of the SVIS ® has now been widely used in laboratories and seed companies in the United States, and results have shown that the procedure is promising for evaluation of the seed vigor of various species (Marcos-Filho et al., 2009;Chiquito et al., 2012;Silva et al., 2012;Kikuti and Marcos-Filho, 2013;Caldeira et al., 2014).
Kikuti and Marcos-Filho (2012) evaluated the effectiveness of the SVIS ® in lettuce seeds and concluded that the growth index and the primary root length provided consistent results similar to other tests commonly used for this species.Silva et al. (2012) studied the SVIS ® for sun hemp seeds (Crotalaria juncea L.) and observed that this software is effective in detecting differences in the physiological potential of seed lots and that, associated with the electrical conductivity test, it can provide important information regarding the seed vigor.Evaluation of seed vigor using the SVIS ® has also been reported as successful for species such as lettuce (Sako et al., 2001 andPeñaloza et al., 2005), melon (Marcos-Filho et al., 2006), soybean (Hoffmaster et al., 2005and Marcos-Filho et al., 2009), eggplant (Silva and Cicero, 2014), and sunflower (Rocha et al., 2015).Thus, procedures effective in discrimination of crambe seed lots would be alternatives for precise identification of seed physiological potential by seed companies.The aim of this study was to evaluate the methodology of the controlled deterioration test and verify the effectiveness of the Seed Vigor Imaging System (SVIS ® ) in evaluation of the physiological potential of crambe seeds.

Material and Methods
The study was conducted at the Seed Analysis Laboratory, Department of Crop Science, Sao Paulo State University (UNESP), Jaboticabal Campus, SP, Brazil, and at the Image Analysis Laboratory, Department of Crop Science, University of Sao Paulo (USP/ESALQ), Piracicaba, SP, Brazil.Five commercial lots of crambe seeds from FMS Brilhante cultivar were used, produced by the Fundação MS, located in Maracajú, MS, Brazil.
Seeds were placed in woven polypropylene bags and kept in a cold chamber (10 ºC and 50-60% relative air humidity -RH) while conducting the laboratory studies.The experiments were set up in a completely randomized design and the following evaluations were made: Seed moisture content: determined by the oven method at 105 ± 3 ºC for 24 h (Brasil, 2009) with two subsamples of approximately 2.0 g of seeds for each lot, before and after the controlled deterioration test.The results were expressed as a percentage on a fresh weight basis.
Germination: four replications of 50 seeds per lot were distributed on 300 g of sieved and sterilized sand in plastic germination boxes (11 x 11 x 3.5 cm) and covered by 0.5 cm of sand.The substrate was previously moistened to 70% of field capacity.To moisten the substrate, a KNO 3 solution (0.2%) was used, according to Brasil (2009).The boxes with the seeds were kept under alternating temperature of 20-30 °C and a 12-h photoperiod.Seedling evaluations were carried out according to the criteria established in the Rules for Seed Testing (Brasil, 2009) at four days for the first germination count, and at seven days after sowing for the final germination count.The results were expressed as a percentage of normal seedlings.
Field seedling emergence: four replications of 50 seeds per lot were sown, arranged in an earth plant bed without fertilization in 2.5-meter length rows, spaced at 0.5 m, at a depth of 2 cm.A spray-irrigation system was used.The mean temperature in the environment while the test was conducted was 23 °C, with a minimum of 15 °C and maximum of 32 °C.Evaluations were made at 14 days after sowing, and only the seedlings that had a hypocotyl longer than 1.0 cm were considered; the percentage of normal emerged seedlings was determined.
Controlled deterioration: the seed moisture content was adjusted to 18%, 20%, and 22% using the wet atmosphere method.For that purpose, 5 g samples of seeds from each lot were distributed on a single layer on a stainless steel mesh inside transparent plastic boxes (11 x 11 x 3.5 cm), which contained 40 mL of deionized water.The boxes were closed and kept in a chamber at 20 ºC, and seeds were weighed periodically until they reached the weight corresponding to the pre-established moisture content.The seed moisture content was adjusted according to the equation: in which: FW = final weight, IW = initial weight, and MCi and MCd correspond to the initial and desired moisture content, respectively.
After obtaining the desired seed moisture content, the samples were placed in hermetic aluminized plastic bags and kept in a cold chamber (10 ºC and 50-60% RH) for five days allowing hygroscopic equilibrium among the seeds.Then, samples were placed in a bain-marie at 45 ºC for 24 h.After this period, the samples were kept at ambient temperature for 30 minutes.The seeds were then subjected to the germination test, according to the methodology described above, with evaluations on the fourth and fifth day after sowing.
Image analysis (SVIS ® ): for analysis using the SVIS ® software, four replications of 25 seeds were used for each lot.The seeds were placed in two horizontal rows on the upper third of the surface of paper towels that was previously moistened with KNO 3 solution (0.2%) and distributed so that growth of the primary root was turned downward.The towels were rolled up and transferred to a germination chamber at 20-30 °C and kept for three days in continuous light.After this period, the seedlings were arranged on a sheet of black pasteboard set on the surface of an aluminum box (60 x 50 x 12 cm), inside of which was an HP Scanjet 2004 scanner, placed upside down and operated by Photosmart software with 98 dpi resolution.
The seedling images were digitalized and then analyzed by the SVIS ® software.The vigor index and length and uniformity of seedling development were evaluated.The vigor index was generated by a combination of growth parameters (70% contribution) and seedling uniformity (30% contribution), both based on the maximum possible length of the crambe seedlings at three days after sowing (3 inches = 7.6 cm).
The data were subjected to analysis of variance by the F test.For the controlled deterioration test, the data were transformed in arc sen √x/100 allowing for homogeneity of variances.The Tukey test at 5% probability was used for comparison the means; data were presented without transformation.In the end, a simple correlation analysis was made among the uniformity of seedling development and the germination test, first germination count, field seedling emergence, and controlled deterioration test.

Results and Discussion
All seed lots were homogeneous regarding to the initial moisture content, with variation of 0.8 percentages points between the extremes (Table 1).Similar results were obtained from the same cultivar of crambe seeds in the studies of Leão et al. (2016).A small variation in moisture content is important because can ensure efficient standardization of procedures for the controlled deterioration test so that seed lots can reach the predetermined moisture content in similar periods (Silva and Vieira, 2010) and maintain a uniform deterioration process (Dutra and Medeiros-Filho, 2008).
In initial characterization of the seed lots (Table 1), the germination percentage of all the lots was acceptable for the commercialization of crambe seeds, which is fixed at 60% (Brasil, 2013), and they indicated lower performance  -------------------------% ---------------------- for lot 5.The first germination count and the field seedling emergence showed differences in vigor among lots with similar germination.Lot 1 was identified as having the highest vigor, exceeding lots 2 and 4 regarding to the seedling emergence, and lot 2 for the first germination count.Lot 5 showed the worst performance in all these physiological potential evaluations.
The moisture content of the five seed lots was adjusted to 18%, 20%, and 22% before the controlled deterioration test was carried out, with relatively small variation among the lots (Table 2).This fact is of great importance for the reliability of the test results because can ensures a similar deterioration process among the seed lots (Dutra and Medeiros-Filho, 2008).As highlighted by Powell and Matthews (1981), one limitation of the controlled deterioration test as routine test includes the precision required for increasing the seed moisture content.
During the controlled deterioration test, greater sensitivity in differentiating physiological potential of the lots was observed when the seed moisture content was adjusted to 18% and evaluation was carried out on the fifth day after sowing (Table 3).These results were similar to the field seedling emergence, which also identified lot 1 as the most vigorous and lot 5 as the lowest vigor.Evaluation of controlled deterioration on the fourth day was able to separate lots into only two distinct physiological classes with results different from those obtained for field seedling emergence, indicating lower sensitivity for detecting differences in vigor among the seed lots.
Similar results, in which adjustment of seed moisture content to values under 20% proved to be promising were obtained for other species.For example, Mendonça et al. (2003) concluded that moisture content adjusted to 18% during the controlled deterioration test is also efficient to evaluate broccoli seed lots vigor.Rossetto et al. ( 2004) evaluated the controlled deterioration test in peanut seeds with moisture contents of 15% and 20% and observed that adjustment to 15% showed high correlation with other seed vigor tests.However, in contrast, Torres et al. ( 2012) studied coriander seeds with adjustments to 18%, 21%, and 24% and found that moisture content adjusted to 21% reveals higher effectiveness in separating seed lots into different vigor levels.
In this study, the controlled deterioration test with initial adjustment of moisture content to 20% and evaluation at four and five days after sowing were in contrast with the other evaluations, classifying lot 1 as low vigor (Table 3).However, this test with adjustment of seed moisture content to 22% had a response nearer to the results obtained during initial evaluations (Table 1).Nevertheless, the seed vigor classification was not similar to the field seedling emergence, which exhibited lower sensitivity in ranking seed lots into vigor levels when confronted with the moisture content adjustment of 18% in evaluation on the fifth day after sowing.
Rossetto and Marcos-Filho (1995) demonstrated that the higher the moisture content of soybean seeds, the more damaging the effect of high temperature during the controlled deterioration test.Thus, this does not allow consistent classification of the lots into different vigor levels.In this context, Kikuti and Marcos-Filho (2008) obtained satisfactory results in cauliflower seeds only when the seed moisture content was adjusted to lower values, such as 20% and 22%; satisfactory results were not possible with adjustment to 24%.After controlled deterioration, the seed moisture content remained near the initially adjusted value (Table 4).This factor is relevant because it indicates the degree of reliability of controlled deterioration test results (Marcos-Filho, 2015).1) as well as the controlled deterioration test with the best procedure for testing in crambe seeds, i.e., adjusting the moisture content to 18% and evaluation on the fifth day after sowing (Table 3).
In contrast, the indexes generated by the SVIS ® for uniformity of seedling development were able to separate the seed lots in an effective manner, as can be observed from the digital images of the seedlings (Figure 1).These results are in agreement with those obtained by Silva et al. ( 2012), who concluded that uniformity of seedling development can provide useful information regarding the degree of deterioration, initial growth potential, and uniformity of seedling emergence in sun hemp.For this parameter, reliable results were also found by Sako et al. (2001), Hoffmaster et al. (2005), and Marcos-Filho et al. (2006), confirming the high potential of the uniformity of seedling development parameter, generated by the SVIS ® , to provide information regarding the vigor of seed lots of various species.
The simple correlation coefficients presented in Table 6 reveal that the results of the field seedling emergence exhibited significant correlation only with uniformity of seedling development (r=0.91).The data of first germination count, for their part, presented significant correlation with those of the controlled deterioration test (r=0.83).The strong correlation with vigor tests widely used for various species, demonstrates the possibility of using information from uniformity of seedling development and the controlled deterioration test to evaluate the crambe seed vigor.
Image analysis of crambe seedlings by SVIS ® (Table 5) allowed efficient identification of seed lots into different vigor levels in a similar manner to other tests.Lot 1 was identified as the highest vigor for all the parameters generated by the software.On the other hand, the seedling length and the vigor index showed no sufficient sensitivity to detect different levels of physiological potential among other seed lots; the best separation of lots occurred only for uniformity of seedling development, which in addition to showing the best performance of lot 1 also highlighted lot 5 as having the lowest potential, as confirmed in the field seedling emergence results.
It is noteworthy that the results presented by use of traditional vigor tests were not able to efficiently separate

Table 2 .
Adjustment of moisture content for the controlled deterioration test at 45 °C for 24 h in five lots of crambe seeds.

Table 3 .
Vigor evaluated by the controlled deterioration test in five lots of crambe seeds, with adjustment of the seed moisture content to 18%, 20%, and 22%, at four and five days after sowing.Mean values followed by the same letter in the column do not differ among themselves by the Tukey test at 5% probability.** Significant at the level of 1% probability.

Table 4 .
Moisture content of five lots of crambe seeds after the controlled deterioration test at 45 °C for 24 h.

Table 5 .
Seedling length (cm), vigor index, and uniformity of seedling development from five lots of crambe seeds obtained by the Seed Vigor Imaging System -SVIS ® .
Mean values followed by the same letter in the column do not differ among themselves by the Tukey test at 5% probability.** Significant at the level of 1% probability.

Table 6
*Significant at the level of 5% probability and ns not significant.