PERFORMANCE OF SWEET POTATO CLONES FOR STARCH AND ETHANOL IN THREE REGIONS OF THE STATE OF SERGIPE , BRAZIL

Sweet potato (Ipomoea batatas L.) is a species that belongs to the family Convolvulaceae, and is originated from Central America and South America. As the growing conditions have great influence on the crop, the determination of harvesting time may vary with the cultivar, the growing region, or with the type of consumption (in natura or industrial). The aim of this work was to evaluate the performance of Ipomoea batatas L. clones, cultivated in three regions in the state of Sergipe, for starch and ethanol production. Thirty-one sweet potato clones grown in three municipalities of the state of Sergipe were tested in a randomized block design. The following variables were analyzed: root dry matter content (RDMC), root starch content (SC), starch yield (SY), ethanol yield (EY), and ethanol yield per ton of root (EYR). EY values ranged from 5910.39 to 8516.12 L ha; from 5141.85 to 6937.63 L ha; and from 5829.62 to 8211.77 L ha in the municipalities of São Cristóvão, Malhador, and Canindé de São Francisco, respectively, for clones IPB-075 and IPB-087 and cultivar Palmas. Estimates of heritability (h) were above 50%. The values of the ratio between the coefficient of genetic variation (CVg) and the coefficient of environmental variation (CVe) for RDMC, SC, and EYR were high.


INTRODUCTION
Sweet potato (Ipomoea batatas L.) is a species that belongs to the family Convolvulaceae; it is originated from Central America and South America, and is grown from the Yucatan Peninsula, in Mexico, to Colombia and Brazil (SILVA et al., 2004).Despite being grown as an annual crop, sweet potato is a perennial, continuous tuber plant, and its natural death only occurs under severe weather conditions, such as frost, or long drought periods.Under ideal growing conditions, harvest begins at 90 days after planting.However, it can also occur between 120 and 150 days after planting, depending on the growing conditions, on the environmental conditions, and on the variety of sweet potato used (SILVA et al., 2002).Since the growing conditions greatly influence the crop, the determination of harvesting time may vary with the cultivar, with the growing region, or with the type of consumption (in natura or industrial).For industrial use, the species can be harvested later, since the primary interest is to obtain large quantities of dry matter and high carbohydrate levels, which results in higher yields for the industrial process (QUEIROGA et al., 2007).
Ethanol has been produced from starch since the 1970s; since then, scientists have been studying the use of starches for fuel, focusing on raw material yields for conversion into ethanol.While attempting to optimize glucose production from sweet potato starch, Ribeiro et al. (2009) observed that using high temperature and pressure was a more efficient method for acid hydrolysis than using water bath.Given that carbohydrates are converted into ethanol during fermentation, Pavlak et al., (2011) evaluated the efficiency of the fermentation process for hydrolyzed sweet potato using three Saccharomyces cerevisiae strains.The authors found that sweet potato used as raw material presented promising levels of viability, yield, and fermentation process efficiency for ethanol production.
With the expansion of renewable energy and biomass sources, this species has become an option for ethanol production in several regions of Brazil for being an easy-to-handle rustic plant with genetic diversity.Cavalcante et al. (2009) evaluated the

MATERIALS AND METHODS
The experiments were carried out in three different regions of the state of Sergipe: Region 1the "Campus Rural da UFS" Research Farm of the Federal University of Sergipe, located in the municipality of São Cristóvão, state of Sergipe, in the central portion of the physiographic region of the coast of the state of Sergipe (lat.10°55'27" S, long.37°12'01" W, at 46 m asl); Region 2 -the Dandara settlement project, located in the municipality of Malhador, state of Sergipe, in the Central portion of the agreste region of the state of Sergipe (lat.10°42'37.4"S, long.37°16'05.8"W, at 35 m asl); and Region 3 -the Jacaré-Curituba settlement project, located in the municipality of Canindé de São Francisco, state of Sergipe, in the North-West region of the state of Sergipe, on the right bank of the São Francisco River (lat.09°38'40.1"S, long.37°37'16.1"W, at 68 m asl).
The clones tested in the three environments were obtained from the Active Germplasm Bank (AGB) of the Federal University of Sergipe, as described in Table 1.The commercial cultivars Brazlândia Branca, Brazlândia Rosada, and Palmas were used as controls.Vegetative branches with eight to 12 internodes (25 cm) were selected for planting from the population of clones of the AGB of the Federal University of Sergipe, properly identified, and transported to the planting site.Prior to planting, branches had been watered to prevent loss of branches due to dehydration.In the experimental area, branches were planted at 10-15 cm depth, and a portion containing three to four nodes was buried, and subsequently watered.The nutrient requirements of the crops were calculated based on physical and chemical soil analyses (Table 1).At 30 days before planting, the municipalities of São Cristóvão and Malhador were limed at 2 T ha -1 , and the plot was fertilized with NPK (16 g plant -1 NPK 6-24-12).After planting, topdressing was applied monthly, using the same formula, but at 12 g plant -1 .The amount of fertilizer varied according to the soil analyses for each site (Table 2), with a dose of 8 g plant -1 at planting, and 6 g plant -1 as monthly topdressing, in the municipality of Canindé de São Francisco.Plowing and disking were carried out for soil preparation.Ridges were made manually with a hoe.Cultural practices were applied when necessary, according to the recommendations for the crops.Sprinkler irrigation was used to meet the water requirements of the crops.Plants were harvested at 180 days after planting.
The following variables were evaluated: Root dry matter content (RDMC): Obtained by placing 10 g of freshly grated root from each plot in a forced air circulation drying oven, at 105°C, until constant weight.RDMC was calculated using the formula: Where RDMC is the root dry matter content, RDM is the weight of root dry matter, and RFM is the weight of root fresh matter; Root starch content (SC): Obtained by the Lane-Eynon method (INSTITUTO ADOLFO LUTZ, 2005).The method is based on reducing a known volume of an alkaline copper reagent (Fehling) to cuprous oxide; Starch yield (SY): Obtained by multiplying the starch content by the total root yield, and expressed in T ha -1 ; Ethanol yield (EY): Obtained by multiplying the starch yield by the conversion factor 1.0858, according to Pavlak et al., (2011), and expressed in L ha -1 ; Ethanol yield per ton of root (EYR): Obtained by dividing the EY by the total root yield, and expressed in L ton -1 ; For each site, the experimental design consisted of randomized blocks with three replications.Each plot consisted of a 0.40 m ridge of 10 plants, spaced 0.80 m between ridges and 0.40 m between plants.IPB-149 clone was planted (local cultivar) as border, surrounding the experiment.
The variables were subjected to joint analysis of variance, according to the model proposed by Vencovsky & Barriga, (1992), considering the effects of environments and genotypes as random, using the software GENES.
The following mathematical model was used: where Y ijk is the observed value of the i th genotype, in the k th block, within the j th environment; m is the overall mean; B/A jk is the effect of the k th block within the j th environment; G i is the effect of the i th genotype; A j is the effect of the j th environment; GA ij is the effect of the interaction between the i th genotype and the j th environment; and E ijk is the experimental error.
Genetic and phenotypic parameters, such as heritability (h 2 ), coefficient of variation (CV), and phenotypic and genotypic correlations were estimated using the GENES software (Cruz 2006).The Scott-Knott test was used for means clustering, at 5% probability, using the SISVAR software, version 5.0.
Based on the expected mean squares, the genetic ( 2 g σ ) and environmental ( 2 e σ ) variance components were estimated for the primary evaluated traits.Genetic parameters were also estimated, which included the coefficient of heritability (h 2 ) and the ratio between the CV of genetic variation and the CV of environmental variation (CV g /CV e ).

RESULTS AND DISCUSSION
Differences between treatments (α = 0.05) were observed for all the evaluated traits: root dry matter content (RDMC), root starch content (SC), starch yield (SY), ethanol yield (EY), and ethanol yield per ton of root (EYR) (Tables 3 and 4).Differences were also observed for joint analysis and genetic parameters (Tables 5 and 6).

Root dry matter content (RDMC)
For the municipality of São Cristóvão, three groups were formed for RDMC, with values ranging from 29.72% to 44.05%.IPB-159 clone had RDMC value (44.05%) higher than those found in the controls Brazlândia Branca and Brazlândia Rosada (Table 3).
For the municipality of Malhador, two groups were formed for RDMC, with values ranging from 24.26% to 36.82% (Table 3).IPB-162 clone had RDMC value (36.82%) similar to that of the control Brazlândia Rosada (Table 3).
For the municipality of Canindé de São Francisco, also two groups were formed for RDMC, with values ranging from 26.74% to 40.48% (Table 3).Clones IPB-162 (38.59%) and IPB-149 (40.48%) had RDMC values higher than those found in the controls Brazlândia Branca and Brazlândia Rosada (Table 3).Comparing the results found in each municipality, RDMC values in São Cristóvão were higher than those found by Silveira (2004), who evaluated nine sweet potato cultivars, and observed a range of RDMC values from 28.56% to 40.44% in cultivars recommended for use in bioenergy production.Vieira et al. (2015) found similar results, with values of RDMC ranging from 25.12% to 37.67%.This shows that different genotypes present different responses to RDMC, depending on the environment where they are grown.In the municipality of Malhador, RDMC values were lower than those found by Silveira (2004), and similar to those found in the municipality of Canindé de São Francisco.

Root starch content (SC)
In the municipality of São Cristóvão, clones had different SC values and formed four distinct groups.Clones IPB-162, IPB-158, IPB-149, and the control cultivar Palmas presented SC values of 23.32%, 24.63%, 25.64%, and 27.26%, respectively (Table 3), which were higher than those found for the other cultivars.These materials have proved to be promising, demonstrating the importance of competitive trials in multiple environments.SC values of the other treatments ranged from 11.47% to 22.60% (Table 3).
Differences were observed among treatments of the municipality of Malhador.Five groups were formed, and IPB-162 (26.06%),IPB-159 (28.58%), and IPB-056 (24.16%) presented the The degree of variability within the measured variables differed within each environment, with small reduction in RDMC in São Cristóvão; small reduction in RDMC and EYR in Malhador; and small reduction in RDMC in Canindé de São Francisco (Table 5).Azevedo et al. ( 2014) observed the influence of the genotype x environment interaction only for starch content.In the present work, this influence was observed for all the studied variables.
For Malhador, the highest CV g values were observed for RDMC.Canindé de São Francisco had the highest CV g values for SC, SY, EY, and EYR, indicating the presence of greater variation for these traits, when compared with the other variables.
In the present experiment, the values of CV g /CV e ratio for RDMC, SC, and EYR were high (0.91, 1.42, and 1.42, respectively), justifying their use in the genetic improvement process (Table 6).For ethanol biofuel production in the three environments, cultivar Palmas and clones IPB-075 and IPB-087 presented the best performance for the following traits: RDMC, SY, and EY.h² estimates were above 50% for all variables, demonstrating that this trait is highly influenced by the environment and is not a good indicator of the genotypic value (Table 6).According to Cruz (2005), high h² values indicate that genetic differences are responsible for the variations in traits, whereas low h² values indicate that much of the variation is due to environmental differences among individuals.High h² values indicate that the traits have strong genetic control and great potential to be transferred to future generations.
All of the traits, except for SY, and EY, presented high heritability, indicating that the traits have strong genetic control and great potential to be transferred to future generations.With this, the best-

Table 1 .
Description of the 31 evaluated sweet potato clones and their identifications in the Active Germplasm

Table 2 .
Results of the chemical and physical soil analysis of the experimental areas of São Cristóvão, Malhador and Canindé do São Francisco, at 0-20 cm depth.

Table 3 .
Mean values for content of root dry matter, root starch and starch yield for 31 sweet potato clones grown in three municipalities of the state of Sergipe, Brazil.Means followed by the same lowercase letter in the columns and uppercase letter in the rows do not significantly differ by the Scott-Knott test (p ≤ 0.05); SC = São Cristóvão; MA = Malhador; CA = Canindé do São Franscisco.

Table 5 .
Summary of the analysis of covariance of the following variables: root dry matter content (RDMC), root starch content (SC), starch yield (SY), ethanol yield (EY), and ethanol yield per ton of root (EYR), for 31 sweet potato clones grown in three municipalities of the state of Sergipe, Brazil.

Table 6 .
Genetic parameters for joint analysis and in each environment: root dry matter content (RDMC), root starch content (SC), starch yield (SY), ethanol (EY), and ethanol yield per ton of root (EYR), for 31 sweet potato clones grown in three municipalities of the state of Sergipe, Brazil.
g : coefficient of genetic variation; CV e : coefficient of environmental variation.