Morpho-agronomic characterization of sweet potato germplasm

Sweet potato (Ipomoea batatas) spreads mainly by vegetative propagation, and therefore this vegetable may have duplicate accessions in germplasm banks. The authors aimed to characterize morpho-agronomically the sweet potato germplasm from the Active Germplasm Bank (AGB) of Universidade Federal de Sergipe (UFS). The experiment was carried out at the Research Farm “Campus Rural da UFS”. The experimental design was a randomized block design, testing 73 sweet potato accessions with two replicates and four plants per plot. Morpho-agronomic traits of aboveground part and roots were evaluated. We observed that the leaf lobe type, the number of lobes per leaf, shape of central lobe, and general leaf shape were traits which provided the most variability among the accessions. For damage caused by soil insects, 52 accessions showed tolerance. The accessions presented a range from 0.33 to 2.71 t/ha for dry mass of aboveground part and from 1.20 to 10.89 t/ha for the total productivity of the roots. The high phenotypic variability of this crop shows good prospect for breeding programs.


Research
Hortic.bras., Brasília,v.35,n.4,October-December 2017 S weet potato (Ipomoea batatas)   is a plant cultivated worldwide, where water availability is enough to support its growth (Salawu & Mukhtar, 2008).Besides the economic and social importance, it is source of food supplement for population living in poverty areas, since it is rich in carbohydrates and also highly energetic (Camargo, 2013).
This vegetable is easy to cultivate, rustic and widely adapted to different types of soil and climate, with high tolerance to dry weather and low production cost.Sweet potato is widely used in human diet and animal feed and as raw material in food, fabric, paper, cosmetics, adhesive preparation and fuel alcohol industries (Gonçalves Neto et al., 2011).
The world's largest producer of sweet potatoes is China, producing 3,685,254 t/year reaching average productivity of 23.1 t/ha (FAO, 2010).Brazil has an annual productivity of 12.8 t/ha, considering that Rio Grande do Sul is the largest producer, with 166,354 t and productivity of 13.4 t/ha.The State of Sergipe contributes with 44,397 t, considering to be the largest producer in the Northeast part of Brazil with productivity of 14.3 t/ha (IBGE, 2013).
The sweet potato shows high genetic variability, however, changes in consumption habit and lack of researches have been contributing to the loss of important genotypes.Due to this fact, the autors justify the importance of maintaining this species in germplasm banks (Figueiredo, 2010).
Morphological markers are very accessible descriptors for quantifying the phenotypic diversity when compared to more advanced molecular techniques, and these markers have been used by several authors in Brazil and in the world, allowing some authors carrying out studies with more sophisticated descriptors, such as the molecular markers (Fabri, 2009).
The agronomic characterization

Received on June 10, 2016; accepted on April 13, 2017
provides information for sweet potato breeding, for commercial cultivation and to determine the use of the accessions.These descriptors take into consideration desirable traits which meet the market demandings.For sweet potato, these traits are total and commercial productivity, adaptability and stability of accessions in a particular growing region, occurrence of skin defects, resistance to pests, among others (Camargo, 2013).
Given the above, the authors aimed to characterize morpho-agronomically the Sweet Potato Germplasm Bank of Universidade Federal de Sergipe.
Liming was carried out thirty days after planting based on the results of the soil analysis.According to Filgueira (2008), fertilization consisted of 20 kg/ha of N; 90 kg/ha of P 2 O 5 ; 90 kg/ha of K 2 O.At 45 days after planting, the authors carried out the top dressing, applying 30 kg/ha of N, 40 kg/ha of P 2 O 5 and 60 kg/ ha of K 2 O.The supplementary cultural management during the conduction of the experiment was carried out according to recommendations for sweet potato (Filgueira, 2008).
The experimental design was randomized blocks, with seventy three treatments and two replicates, with four plants per plot, spacing used was 1.0 x 0.3 m.The authors used vines containing from eight to twelve internodes, 25-cm length, burying the basal part of the vine, containing three-four internodes, at 10-15 cm depth in windrows of 40-cm high.
The morphological characterization of the above ground part was carried out at 120 days after the vine planting.The vegetative parts were evaluated using the following 15 descriptors (Huamán, 1992): growth habit, internode length, predominant vine color, secondary vine color, general leaf shape, leaf lobe types, leaf lobe number, general leaf shape, central lobe shape, mature leaf size, leaf vein pigmentation, mature leaf color, immature leaf color, petiole pigmentation and petiole length.
T h e r o o t m o r p h o l o g i c a l characterization was carried out at 150 days after the planting of vines, during the harvest.The roots were evaluated, according to Huamán (1992), using 7 descriptors: root shape, root surface defects, predominant pellicle color, predominant pellicle color intensity, secondary pellicle color, predominant flesh color and secondary flesh color.
The agronomic characterization was carried out using the descriptors: shoot fresh mass, in which the vines were cut, identified and weighed, being the results expressed in t/ha.After weighing, the vines were kept in paper bags and placed in oven with forced air circulation at a temperature of 65°C until constant weight, the results were expressed in t/ha.The total productivity of the roots was obtained weighing all the roots of the plot and having their productivity estimated in t/ha.
The classification of sweet potatoes in relation to damage caused by soil insects was carried out using the rating scale, established by França & Ritschel (2002), ranging from 1 to 5, in which "1" represents damage-free roots, suitable for market; "2" roots showing little damage, losing a little regarding to the commercial aspect (presence of some galleries and holes in the roots); "3" roots showing damages which are easy to be noticed (presence of some galleries and holes in the roots in greater intensity), unsuitable for market; "4" highly damaged roots, practically unsuitable for the market (presence of many galleries, and early rot holes); and "5" roots completely unsuitable for commercial purposes (full of galleries, holes and in an advanced stage of rotting).
Data were submitted to variance analysis and the averages compared by Scott-Knott test p≤0.05,using the statistical program SISVAR (Ferreira, 2011).To estimate the phenotypic diversity, multivariate analysis was carried out using principal component Morpho-agronomic characterization of sweet potato germplasm analysis using the computer program STATISTICA version 7.0.

RESULTS AND DISCUSSION
The authors observed differences for the evaluated accessions regarding all morphological variables, both for shoot and root (p≤0.01).
The average of the morphological characterization data showed that a variation among twenty one qualitative traits of sweet potato was noticed.The variables which showed little morphological variability were: mature leaf size, 80.8% of the accessions presented average size of 8-15 cm, mature leaf color, which presented yellowish-green color in 83.9% of the accessions, secondary pellicle color, 78.1% of the accessions presenting absence of color, and secondary flesh color, 97.3% of absent-color accessions.
For growth habit, 15.1% of the accessions showed erect vine, 41.1% semi-erect vine, 39.7% spreading and 4.1% extremely spreading.In relation to the internode length, 23.3% resulted in very short internode, 56.2% short, 19.2% intermediate and only 1.3% showed to be long.The predominant color of green vine appeared in 6.9% of the accessions, 41.1% of the accessions showed green vines with few purple spots, 15.1% green vines with many purple spots, 8.2% green vines with many dark purple spots, 20.6% mostly purple vines, 2.7% mostly dark purple vines, 4.1% totally purple vines and 1.3% totally dark purple vines.For the secondary vine color, the authors observed 20.6% with green tip, 34.2% with purple base, 4.1% with purple tip, 27.4% with purple internodes and 13.7% showed other colors.The authors also verified that 27.4% of the accessions showed cordate leaves, 49.3% triangular leaves, 4.1% hastate leaves and 19.2% lobed leaves.In relation to the leaf lobe type, 68.5% of the accessions showed lateral lobes, 6.8% showed moderate lobes and 5.5% showed deeb lobes.For the number of leaf lobes, 72.6% of the accessions showed one lobe, 8.2% three lobes and 19.2% five lobes.In relation to the shape of central leaf lobe 57.6%,The skin showed to be predominant white (9.6%), cream (20.6%), yellow (1.3%), orange (2.7%), brownish orange (19.2%), pink (26.1%), purple red (9.6%) and dark purple (10.9%).The intensity of the skin color was pale in 34.3% of the accessions, intermediate in 30.1% and dark in 35.6%.The secondary skin color was absent for 78.2% of the accessions, cream (5.5%), orange (2.7%), brownish orange (6.8%) and pink (6.8%).The flesh color was mostly white in 30.1% of the accessions, cream in 41.1%, dark cream in 13.8%, pale yellow in 2.7%, pale orange in 6.8%, and strongly pigmented with anthocyanins in 5.5%.In relation to secondary flesh color, about 97.4% was absent, 1.3% was purple and 1.3% dark purple.The authors also noticed that 6.8% of the accessions did not show deffects on the root surface, 45.2% showed alligator-like skin, 24.7% showed veins and 23.3% showed horizontal constrictions.
The morphological variability of the results obtained through hierarchical clustering analysis method of average distance between clusters of the accessions highlighted the existence of phenotypic variability.The authors observed two clusters, one with 16 Averages followed by the same lowercase letters in columns did not differ among them, by Scott-Knott test (p≤0.05).* Notes from 1 to 5, where "1" represents damage-free roots and 5 totally unsuitable roots.accessions (21.92%) and the other cluster composed of 57 accessions (78.08%).The accessions IPB-001 and IPB-155 were the most divergent; on the other hand, the slightest divergence was observed in accessions IPB-095 and IPB-063 (Figure 1).The authors could observe a great variability, without duplication of the accessions.The results found by Moulin et al. (2014) in Campos do Goytacazes-RJ identified five clusters with accessions morphologically duplicated; this was not observed in this study.When the authors characterized morphologically thirteen sweet potato accessions of the Active Bank of Embrapa Temperate Climate of Federal University (Pelotas), the formation of four clusters with subdivisions was verified (Silva et al., 2009).
The eigenvectors values show, in module, the order of importance and the contribution of the morphological descriptors for the variability of the studied group.Thus, in relation to descriptors used, the authors can state that for the first component (Y1), the most discriminating descriptors were: leaf lobe type (-0.919828), leaf lobe number (-0.904024), shape of central lobe (-0.856486) and general leaf shape (-0.839520).In relation to the order of importance of the descriptors in the second principal component (Y2), the authors observed a change in contribution order.
Dispersion graph was consistent with the values of eigenvectors for the first two principal components.The principal components analysis model with two principal components, explained 28.07% of the total variation.The primary principal component showing 17.39% and the secondary showing 10.68%, separated variables leaf lobe type (TLF), leaf lobe number (NLF), central leaf lobe shape (FLC) and general leaf shape (FGF) of the other variables (Figure 2).
The results found in this study were similar to the ones observed by Camargo (2013) in which the traits leaf lobe type (TLF), leaf lobe number (NLF), central leaf lobe shape (FLC) and general leaf shape (FGF), showed the highest values for relative contribution to phenotypic divergence of 40 sweet potato plant introductions.However, Elameen et al. (2011) observed that the traits which contributed the most for genetic divergence of 105 sweet potato accessions were root surface deffects, secondary pellicle (skin) color and secondary flesh color.
The information obtained through the principal component analysis allows identifying the most discriminating and informative morphological descriptors concerning studies on morphological characterization, allowing further studies related to morphological characterization of Ipomoea batatas, reducing and optimizing time for analyses (Yada et al., 2010).
For damages caused by soil insects, accessions showing lower notes are desirable, since they represent damagefree roots or roots showing little damage, suitable for market.Difference for this trait was observed through F test.The authors noticed a great variability in relation to the resistance of the genotypes to damage caused by soil insect, in which 52 accessions showed to be tolerant to insect attack and 21 were characterized as roots which are unsuitable for market.The accessions which showed the best averages can be used as donors of alleles which confer resistance to soil insects in breeding programs (Table 1).
Euscepes postfasciatus, known as West Indian sweet potato weevil, causes damage on the leaf surface, root and flesh, making the product unfeasible for commercialization (Bottega et al., 2009).When Silva (2010) evaluated 100 experimental accessions from botanical seeds of Universidade Federal do Tocantins, found an average value of 2.0 (roots showing little damage, losing a little regarding to the commercial aspect) degree of tolerance to soil pests for the accessions BDGU#84, BDGU#73 and Ana Clara cultivar.Higher average than the one found in

Table 1 .
Damages caused by soil insects, fresh and dry masses of the shoot and total root productivity of 73 sweet potato accessions from the Active Germplasm Bank of Universidade Federal de Sergipe.São Cristóvão, UFS, 2012.