Breaking dormancy, alternative substrates, and the production of Samanea tubulosa (Benth.) seedlings

We investigated different dormancy breaking methods and substrates for Samanea tubulosa seedlings. Four substrate formulations were tested: SCO: commercial substrate; BBA: decomposed babassu biomass + soil (1:1); BBU: buriti biomass + soil (1:1); and BCA: carnauba bagana + soil (1:1), and three dormancy breaking methods: EM: mechanical scarification; IA: immersion in water; and EQ: chemical scarification. The highest seedling emergence was in BCA substrate with chemical scarification. BCA and SCO provided the highest average seedling height, root length, root volume, root density, diameter at neck height, number of leaves, aboveground dry mass, belowground dry mass, and Dickson quality index values. Therefore, BCA can be used as a substrate for the production of S. tubulosa seedlings, and the immersion in water method is optimal when combined with a BBU substrate.


INTRODUCTION
Unsustainable agricultural practices for land use and management conducted by farmers in the humid tropics, such as along the periphery of the Amazon in the Brazilian state of Maranhão, are factors related [1] that increase soil degradation [2]. In addition, the practice of itinerant cutting and burning has accelerated deforestation in the Amazon [3,4].
The use of native species is an important alternative in the recovery of degraded areas, as it allows the protection and enrichment of the soil, the control of erosion, less loss of water through evaporation and greater biological activity [5]. Among these species, Samanea tubulosa (Benth.) Barneby & JW Grimes has been used in the restoration and maintenance of ecosystems in degraded areas [6,7], being characterized as a typical tree in the Cerrado biome, reaching up to 28 m in height and 1 m in diameter, widely used in forestry, ornamentation and as part of animal nutrition [8,9].
According to Dutra et al. (2013) [10], mechanical scarification, chemical scarification (usually with sulfuric acid), and hot water immersion are the notable methods for breaking the dormancy of seeds of forest species. These methods aim to dissolve the waxy cuticular layer, reducing the hardness of the integument and allowing an accelerated and uniform seed germination process.
Seed dormancy is one of the main obstacles in the production of S. tubulosa seedlings, and results in seed loss, longer production times, and seedling unevenness [11]. Dormancy consists of delaying seed germination, which guarantees, from an ecological point of view, the distribution in time and space, however, for seedling production the attribute is not desirable, as there may be a low percentage of germination [12].
The use of commercial substrates that are based on expanded vermiculite and organic materials [13] is costly [7,14], however, decomposed organic residues and fractions from Brazilian pre-Amazonian palms, such as babassu (Attalea speciosa Mart.), buriti (Mauritia vinifera Mart.), and carnauba (Copernicia prunifera Mill.), can reduce production costs without compromising seedling quality [14,15,16,17,18]. The aim of this study was to investigate three dormancy breaking methods (namely, mechanical scarification, immersion in water, and chemical scarification) and different substrates for S. tubulosa seedling production.

Location and climate
The experiment was conducted from April to June 2018 at the Center for Agricultural and Environmental Sciences (CCAA) of the Federal University of Maranhão (UFMA), Chapadinha, Maranhão, Brazil (3°44ʹ12.62′′S, 43°19ʹ03.51′′E, and 105 m above sea level). The region's climate is classified as humid tropical, with an average annual rainfall of 1613.2 mm and an average annual temperature of 26.9 °C [19].

Experimental design, treatments, and characterization procedures
The study was conducted with a completely randomized design in a 4 × 3 factorial arrangement (four substrates and three dormancy breaking methods) with five replicates containing five plants each, totaling 25 plants per treatment and 300 plants in total, corresponding to 60 experimental units. Four substrate formulations were tested: SCO: commercial substrate; BBA: decomposed babassu biomass + soil (1:1); BBU: buriti biomass + soil (1:1); and BCA: carnauba bagana + soil (1:1), and three dormancy breaking methods: EM: mechanical scarification; IA: immersion in water; and EQ: chemical scarification.
SCO was composed of pine bark and vermiculite (Plantmax HT ® ) and was used as a control, and soil used in the formulation of BBA, BBU, and BCA was collected from the 0-20 cm soil layer in the UFMA/CCAA experimental area. The soil was classified as a dystrophic Yellow Latosol [20], contained 560, 190, and 250 g kg -1 of sand, silt, and clay, respectively, had an "average" textural classification, and a degree of flocculation of 77 g 100 g -1 .
The soil and substrate samples were sent to the Soil Science Laboratory of the Federal University of Ceará, Fortaleza, to measure global density, particle density, and total porosity ( Table  1) according to the procedures described by Schmitz et al. (2002) [21], as well as pH, organic matter content, nitrogen, phosphorus, potassium, calcium, magnesium, and sulfur (Table 2) by the methodology described by MAPA (2007) [22].  The S. tubulosa seeds were purchased from local stores. EM was performed using No. 180 sandpaper until the shiny surface of the seed coat was removed, IA was conducted in distilled water at room temperature for 24 h, and EQ was performed using concentrated sulfuric acid (H2SO4) for 5 min, followed by washing for 5 min in running water.

Experimental procedure
Seeds were distributed in polyethylene bags (12 × 20 cm) with two seeds per bag, in which they were buried at a depth of 1 cm for a period of 60 days. The bags were placed on a wooden bench in a greenhouse under controlled conditions. Daily watering was carried out to ensure adequate moisture within the substrate.
The count of emerged seeds was performed until 30 days after sowing (DAS), considering the seeds that contained the cotyledons fully open above the substrate. Subsequently, thinning was done, leaving only one plant per bag. The emergency percentage (%E) was calculated based on these data. At 60 DAS were measured: seedling height (SH, measured from the neck to the insertion of the last leaf), root length (RL), diameter at neck height (DNH), number of leaves (NL), root density (RD), aboveground dry mass (ADM), belowground dry mass (BDM), and total dry mass (TDM).
For the height and length measurements, rulers graduated in mm were used, for the diameter, a digital caliper was used, and to determine ADM and BDM, fresh material was placed in a forcedair-circulation oven for 72 h and then weighed on a precision scale. The Dickson quality index (DQI) was determined according to Equation 1 [23].
Where Yeij is the response variable, "µ" is the overall average, Mi is the i-th method of breaking dormancy, Sj is the i-th substrate, Mi × Sj is the interaction between the Mi-th dormancy breaking method and the Sj-th substrate, and Erroij is the random error.
Where x and y are the sample means of x and y, respectively, and Sx and Sy are the sample standard deviations of x and y, respectively.

RESULTS
There was a significant interaction effect between substrate type and dormancy breaking method on seedling emergence (Table 3). Using SCO, EM resulted in the highest seedling emergence (70.07%). BCA and BBA provided the highest seedling emergence using EQ, with respective values of 78.08 and 56.05%. Using BBU, the best method for breaking dormancy was IA, with 50.05% seedling emergence. There was no significant interaction effect between dormancy breaking method and substrate (P > 0.05) on SH, RL, root volume (RV), or RD, suggesting that these factors acted independently (Table 4). .44 cm, respectively, which did not significantly differ (P > 0.05) ( Figure 1A). Seedling RLs in BCA and BBU were similar to those in the control (SCO) (Figure 1B), and SCO (2.63 mL) and BCA (2.11 mL) had the highest average RVs, which did not significantly differ ( Figure 1C). BCA (0.17 g mL -1 ), SCO (0.15 g mL -1 ), and BBU (0.15 g mL -1 ) gave the best results for RD ( Figure 1D) and did not significantly differ. The BBA-based substrate gave the lowest average SH, RL, RV, and RD values. There was a significant interaction effect between substrate type and dormancy breaking method on DNH, NL, ADM, BDM, and DQI (Table 5). SCO and BCA provided the highest average DNH, regardless of the dormancy breaking method employed. There was no difference between methods using the BBA-based substrate, which had the lowest averages. The BBU substrate gave higher DNH using the EQ method than EM (Table 5). The use of SCO, BCA, and BBU resulted in the highest NL, with no differences between the dormancy breaking methods. Using BBA, EM provided a higher NL than AI (Table 5). There were no significant differences in ADM among the methods, regardless of the type of substrate used, except for BCA, which provided higher ADM using the AI method than EQ. BCA had the highest values, regardless of the dormancy breaking method employed.
The highest BDM values were found using SCO and BCA, which were not significantly different. Using BCA, the best methods were EM and IA, while using BBU, the best dormancy breaking method was EQ. SCO and BCA gave the highest DQI values and did not significantly differ, regardless of the method employed (Table 5). BBA and BBU had the lowest values, and only by using the EQ method was the BBU substrate as successful as SCO and BCA.
We found significant, positive correlations (P < 0.05 and P < 0.01) among all of the variables under study (Table 6), in contrast to Gavioli et al. (2019) [24], who only reported weak correlations among them.
SH was very strongly correlated with ADM and TDM and strongly correlated with RL, RV, BDM, DQI, and RD. NL was only weakly correlated with the other variables ( Table 6).

DISCUSSION
Choosing what type of substrate to use is associated with the inherent physical and chemical characteristics of each substrate, such as porosity and density that influence water absorption and retention, as well as the availability of nutrients necessary for good seedling development Dorneller et al. (2014) [26]. In the present study, the type of substrate used for the production of S. tubulosa seedlings influenced the efficiency of the method for breaking seed dormancy.
EM is a simple, low-cost technique that cracks seeds in order to reduce the impermeability of the seed coat. However, the technique can damage the embryo, and affect germination and seedling emergence [27,28]. In the present study, this method was particularly efficient in SCO by exposing the embryo and facilitating the imbibition process. Although this substrate had the lowest porosity (Table 1), suggesting lower infiltration, the amount of water was sufficient to achieve 70% seedling emergence.
Sulfuric acid accelerates the imbibition process by breaking the seed coat, which results in greater seedling emergence [29]. Using EQ, seedling emergence in BCA and BBA may have been related to their physical characteristics, because their total porosity was 14.03 and 9.13% higher, respectively, than soil with no organic residue (Table 1). In addition, according to Pagliarini et al. (2012) [30], the increase in porosity is associated with an increase in aeration and water infiltration, providing adequate moisture for the emergence of seedlings. Oliveira et al. (2012) [29] investigated pre-germinative treatments for S. tubulosa seeds using washed sand (high porosity) as a substrate, and found that the highest percentage (>70%) of seedling emergence occurred with the use of concentrated sulfuric acid for 5 and 15 min, which is similar to the results of the present study.
IE for 24 h was most effective in BBU, which had >50% seedling emergence probably because of its highly porous nature (63.41%). This provided greater infiltration, optimized seed soaking, and weakened the integument. In this respect, BBU stands out from the other substrates studied, because in addition to being found in abundance in the Amazon and pre-Amazon regions, AI can be used to break dormancy in seeds with hard integuments. This would bring great economic benefits to the seedling producer because the raw material is widely available and cheap, in contrast to the sulfuric acid used in EQ, which has restricted use and is expensive, in addition to harming the environment. The IA method increases the permeability of the seed tissue owing to wear or softening, thus facilitating the diffusion of water into the embryo [31].
Although BCA only had low levels of nitrogen in its composition (Table 2), it contained high levels of organic matter that may have provided a controlled release of nitrogen during seedling development and affected their size ( Figure 1A). According to Guerrini and Trigueiro (2004) [32], organic matter is a fundamental component in the substrates. It is capable of increasing the water retention capacity, improving the physical characteristics of the substrate, and providing nutrients for the seedlings, in addition to being closely related to organic carbon.
BBA gave the lowest average RL ( Figure 1B) probably because of its low levels of phosphorus (Table 2), which is closely linked to root growth. The highest averages for RV and RD ( Figures 1C  and 1D) were attributed to a large root mass, given the very strong and strong significant correlations, respectively, between these variables ( Table 6). The high phosphorous levels in BCA, SCO, and BBU probably increased RD.
SCO and BCA gave the best DNH results, probably because of its strong correlation with RV ( Table 6), suggesting that increased root growth results in greater stem thickening.
NL is an important parameter that is used in choosing the best seedlings in terms of photosynthetic capacity [33]. The results obtained using SCO, BCA, and BBU can be attributed to the levels of nutrients present in each substrate, particularly nitrogen and potassium that are needed in cell division and directly affect seedling growth and development, which are reflected in the NL.
The NL and ADM results were related to the levels of nitrogen and phosphorus present in the BCA substrate (Table 2). According to Faria et al. (2016) [34], these are the most important nutrients during the initial phase of seedling growth, because they play important roles in enzymatic processes, cell maintenance and turgidity, energy storage, and cell growth and elongation. Organic matter is an important component of BCA because it improves the physical, chemical, and biological properties of the substrate.
The high BDM values of plants grown in SCO and BCA can be attributed to their phosphorus concentrations, given their effects on root development, as well as their physical characteristics such as density and porosity, because high porosity and low density provide greater water retention and favor root development [35].
SCO and BCA had macronutrient levels that were adequate for plant development, as reflected in their high DQI values (Table 5). DQI is a measure of seedling quality, and seedlings with a high DQI value have high phytomass, vigor, and, consequently, greater production capacity [36].
SH was very strongly correlated with ADM and TDM ( Table 5), indicating that seedling size is positively related to ADM and TDM values. This may have been due to greater light and carbon dioxide capture, resulting in higher photosynthetic activity.
The very strong correlations found between DQI and ADM, BDM, and TDM were probably caused by the dependence of this index on TDM and the ADM/BDM ratio used in calculating its value. Therefore, S. tubulosa seedlings with high TDM values and good ADM/BDM ratios had high DQI values.
NL was weakly correlated with all of the growth variables (SH, RL, DNH, RV, ADM, BDM, and TDM) and seedling quality (DQI), indicating that NL alone cannot explain the responses of the other variables studied [33].

CONCLUSIONS
Carnauba bagana may be used as an alternative substrate to the commercial substrate for the production of S. tubulosa (Benth.) seedlings; when combined with the chemical scarification method using sulfuric acid, it accelerated the emergence of seedlings. When buriti biomass was used as the substrate, the method of immersion in water for 24 h was satisfactory for overcoming the tegumentary dormancy of S. tubulosa seeds. The use of babassu biomass as the substrate was not suitable for the production of S. tubulosa seedlings.

ACKNOWLEDGMENTS
We would like to thank Editage (www.editage.com) for English language editing.