Greenhouse and Field Evaluation of Selected Sweetpotato ( Ipomoea batatas ( L . ) LAM ) Accessions for Drought Tolerance in South Africa

Sweetpotato is a major staple food in the world. It is a good source of carbohydrates, vitamin A, vitamin C and protein. The maximum production potential of the crop is being hampered by severe drought which ravages most parts of Africa. The main aim of this study therefore was to screen accessions of sweetpotato for drought tolerance in a quick screening method, followed by field screening with a view to identify accessions that can perform well under water stress conditions. Fifty sweetpotato accessions consisting of cultivars and breeding lines collected from the ARC-VOPI gene bank were planted for drought screening in the glass house for six weeks during which water was withheld to induce stress. Observations were made on number of dead plants and days to wilting point. The results were analyzed and 12 best performing accessions were selected for field trials. The field trial was carried out in Lwamondo, Limpopo province, a drought prone area in South Africa, under rain-fed conditions. The best performing accessions were Zapallo, Tacna, Ejumula, 2004-9-2 and Ndou.


Introduction
Rainfed agriculture covers 80% of the world's cultivated land, and is responsible for about 60% of crop production [1].Drought, amongst other environmental factors, contributes to about 75% yield loss each year in the world [2].Agriculture is the largest consumer of water in the world and in the drier areas of the world including South Africa which is regarded a water-stressed country [3].The use of water for agriculture can exceed 90 percent of consumption [4].Global warming is also predicted to affect developing countries severely where agricultural systems are most vulnerable to climatic conditions.[5] predicted considerable increases in number of hot days due to climate change in Limpopo Province South Africa, a popular sweetpotato growing region which will be detrimental to productivity.
The Food and Agricultural Organization of the United Nations [2] estimates that by 2025 approximately 480 million people in Africa could be living in areas with very scarce water, and that as [1] climatic conditions deteriorate, 600,000 km 2 currently classed as moderately constrained will become severely limited.It is thus essential to improve water use efficiency in agriculture.This will require an integrated approach to water resources management to encourage an efficient and equitable use of the resource, and to ensure sustainability.The identification of crop varieties with increased tolerance to drought is therefore an important strategy to meet global food demands with less water.According to [2] the inadequate knowledge of varieties and environmental conditions is causing drought to have such a high impact on yield.
Sweetpotato (Ipomoea batatas) is a major staple food in Africa, Asia, the Caribbean, and South America, where they are important sources of carbohydrates, vitamins A and C, fiber, iron, potassium, and protein [6].Sweetpotato is also used as animal feed.Increasing recognition of the great potential of the sweetpotato crop as a nutritious food for humans and animals has resulted in intensified research efforts to enhance production and consumption in recent decades [7].It has the advantage over other crops in the fact of providing a quick yield during a short growth period and providing a stable productivity less affected by climatic factors [8].Sweetpotato also has flexible planting and harvesting times, tolerates high temperatures and low fertility soils.It is drought tolerant and easy to propagate.Furthermore, compared to other crops, sweetpotato requires fewer inputs and labour making it particularly suitable for households threatened by migration or diseases such as HIV/AIDS [9].The crop is grown over a broad range of environments and cultural practices and is commonly grown in low-input agriculture systems [10].
Sweetpotato, despite being a hardy crop by nature, is sensitive to water deficits particularly during the establishment period including vine development and storage root initiation [11].Drought often is a major environmental constraint for sweetpotato production in areas where it is grown under rain fed conditions [12].However, different genotypes may respond differently to limited quantities of soil water [12] [13].Selection for good cultivar performance under drought conditions is thus considered to be of major importance.Drought necessitates additional irrigation periods, and this increases the overhead production costs.South Africa is regarded as being susceptible to water stress conditions due to prevalence of drought [14].Research into drought tolerant plants is being intensified in order to minimize its overall impact on the agricultural enterprise.The development of drought tolerant sweetpotato varieties will increase profitability to farmers by potentially limiting irrigation and associated production overhead costs.
Genotypic variability exists for drought tolerance with some clones performing better under drought conditions [13].Selection and improvement of adapted genotypes for a particular environment can therefore, be done with the appropriate equipment and using selection criteria associated with drought tolerance [13].With this genetic management option, drought-tolerant varieties, once developed, would be a low economic input technology that may be readily acceptable by resource-poor, rain fed, small land holding farmers.Consequently, the objectives of this study is to screen breeding lines and land races for drought tolerance in a quick screening method, evaluate the selected breeding lines and land races for drought tolerance under field condition and identify accessions that can perform well under water stress conditions without a significant loss of yield and quality.

Pre-Screening
The pre-screening was carried out in the drought screening glass house at Agricultural Research Council-Vege-table and Ornamental Plant Institute (ARC-VOPI), South Africa (Latitude 25.604˚S, Longitude 28.345˚E and Altitude 1159 m).The plant materials used for this experiment were 50 sweetpotato accessions obtained from the gene bank of ARC-VOPI (Table 1).The materials consisted of local cultivars, imported accessions and ARC breeding lines/accessions.The pre-screening of accessions was done between March and August 2008.Stem cuttings of the 50 accessions were planted in plastic boxes of size 155 cm × 77 cm × 23 cm.The boxes were filled with a special soil mixture (5:2:2 sand:soil:vermiculite).Sweetpotato cuttings of about 30 cm long from each accession were cut and planted 2 eyes/nodes below the surface and 3 eyes/nodes above the surface for uniformity of development.Eight accessions were planted in each box with the plant spacing 15 cm between rows and 10 cm between plants and a box contained ten rows (5 plants/row).In each box 8 accessions and 2 control accessions were planted to serve as positive and negative controls namely Lethlabula (drought tolerant and positive control) and Resisto (drought sensitive and negative control).The design was randomized complete block design (RCBD) with 6 replicates consisting of a total of 38 boxes.The experiment was watered for 10 -14 days for establishment after which water was withheld to induce stress.The experiment was concluded 60 DAP when 60% -70% of the plants showed severe stress and wilting.

Field Experiment
The field trial was done at Lwamondo Agricultural Station in Thohoyandou in Limpopo Province, South Africa between March and September 2009.It lies at 23.06'S latitude and 30.38'E longitude with an altitude of 618 m above sea level.Climate conditions are subtropical with average annual rainfall of 752 mm with over 80% occurring between October and December.The average maximum and minimum temperature in the area is between 28.5˚C and 13.7˚C.
The 12 accessions used for the field trials were Zapallo, Resisto, Ejumula, Tacna, 1999-3-1, 2004-16-1, 2004-5-2, 2004-9-2, W-119, Phala, 2003-24-2 and Ndou.These consisted of 10 accessions selected based on the results of the pre-screening trial (Table 2), Resisto, the drought sensitive negative control and Ejumula which was an addition from the previous box experiment.Ejumula was found in previous experiment to be drought tolerant and was therefore used as a replacement for cultivar Japon Tresmesino Selecto which did not have enough planting material available for the field trial.An area of land measuring 20 m × 60 m was used for the trial.Field preparation included making ridges of 0.3 m high and spaced 1 m between the center of the ridges.Before planting Limestone Ammonium Nitrate (LAN, 28% N) was applied at 150 kg/ha (110 g/plot) and super grow (18.5% P) was applied at 150 kg/ha (110 g/plot) to the field by broadcasting method and incorporated into the soil.The cuttings were planted the next day in a triple row of 8 plants per row (24 plants/plot) with spacing of 1 m between rows and 0.3 m between plants and replicated 5 times.The design was randomized complete block design (RCBD).Two border rows were planted on each side of each block.The whole plot was watered using overhead irrigation for 7 days to facilitate plant establishment after which water was withheld till the end of the experiment.Weeding was done manually to remove unwanted plants.

Plastic Box Pre-Screening
Visual observations were made weekly on the condition of the plants.The date was recorded when the plants in a plot showed severe wilting.At the end of the experiment (60 DAP) the number of dead plants per plot were also counted and recorded.The number of days to severe wilting was then calculated.Data analysis for tolerance parameters such as number of days to death and number of dead plants were performed with [15] and included an analysis of variance (ANOVA) to obtain mean values, and the student's protected t-LSD test was calculated at the 0.01 probability level.The multiple t-distribution test procedure [16] was performed to group the lines as sensitive, intermediate or tolerant to drought stress.

Field Experiment
Data were collected on two plants at the middle rows at 42 days after planting (DAP), 84 DAP and 120 DAP and the following parameters were recorded and evaluated.

Plant Growth
Lengths of the shoots were measured using a meter tape and two plants were selected from each plot and meas- ured with the meter tape and the lengths of the shoots were recorded.

Dry Matter Content
Plants were harvested and the fresh weight of both the roots and the shoots were taken separately using a measuring scale.The fresh roots were washed of soil particles and weighed for fresh weight and then cut into pieces and put in a labeled envelope and then oven-dried at 72˚C for 24h00 to get the dry weight.The shoots were placed in a labeled envelope and oven-dried at 40˚C for 24h00 to get the dry weight.The weights were recorded in a data sheet.The dry matter was calculated as follows: ( ) Dry matter % Fresh weight dry weight Fresh weight 100

Canopy Temperature
Recordings of the temperature of the canopy were done in the early hours of the morning using an infra-red thermometer (Raytek Raynger ST20).This was taken at 1 m from the plot edge and 50 cm above the canopy, focusing on the leaves only to reflect exact reading.

Yield
The storage roots of the whole plot were harvested at 120 DAP.The storage roots were thereafter graded into marketable, unmarketable based on their shapes, sizes, weights and defects.Marketable roots were greater than 100 g with a diameter of 3 cm and above when measured and no noticeable pest attack or diseases.Unmarketable roots those less than 100 g and the diameter is less than 3 cm and those showing defects (cracks, irregular shape) and pest infestation.

Data Analysis
Data collected were subjected to statistical analysis using computer software [17].Analysis of variance (ANOVA) was computed and means were compared by the least significance difference at 1% probability level.

Drought Response during Pre-Screening
The results of the pre-screening trial are shown in Table 2.The entries survived between 46 and 60 days before dying.The mean number of days to death was 56.44.The accession with the shortest days to wilting was Yan Shu1 (46.57) and those with the longest day to wilting were Atacama and 2004-16-1 (60.00).Most of the accessions started wilting at 57 days when the experiment was almost concluded indicating that the accessions are generally tolerant to water stress.According to [18], the physiological responses of plants to a deficit of water include leaf wilting, a reduction in leaf area, leaf abscission, and the stimulation of root growth by directing nu-  trients to the underground parts of the plants.Observations made on the number of dead plants after the experiment were a better measure to discriminate the entries in terms of drought response.The least number of dead plants were found in Atacama and 2004-16-1, namely zero.In contrast entries at the bottom of the table had lost 4 plants during drought stress and were very sensitive to drought stress.The best performing accessions that were selected for field trials had potential drought tolerance capabilities which allow them to still flourish under prolong water stress.In addition, imported varieties which performed well in similar pre-screening experiments were also selected.It is expected that these accessions will also show acceptable yield when subjected to periodic moisture stress under field condition.Success in breeding for drought tolerance has not been as pronounced as for many other traits.This is partly due to lack of simple, cheap, and reliable screening methods to select drought-tolerant plants and progenies from the segregating populations and partly due to the complexity of factors involved in drought tolerance.[19] described a simple wooden box pre-screening method showing good correlation with drought tolerance at vegetative and reproductive stages, to select drought-tolerant plants at the seedling stage.Several experiments on drought screening under greenhouse conditions have also been reported in many crops [20]- [24].Wooden box seedling screening is suitable screening large numbers of accessions.It is essential in its ability to determine stress at developmental stages.[19] applied this method in screening large number of cowpea accessions.The parameters used for evaluation were moisture content, flowering, yield and wilting point.
The wooden box method was recommended because it is simple and non-destructive and it can easily be used to screen large number of accessions.In previous experiments, accessions were screened for drought in rain-out shelters [25].Drought tolerance was measured in terms of yield reduction and drought sensitivity index (DSI) of [26].The results of the screening in plastic boxes and planting on the field indicated a relation between number of days to wilting and DSI.These results therefore show that the boxes method is quick, simple and reliable and can be very effective especially for screening large number of accessions.[27] also conducted experiments in rainout shelters on sweetpotato to determine the effect of prolonged restriction of water on yield; accessions were planted in a rain out shelter and irrigated using different water regimes.It was concluded that there was a significant reduction in marketable storage yield with the best yield coming from the accessions supplied with higher volume of water.In general in vitro screening method proves to be an ideal method to screen large set of germplasm with less effort, accurately and the growth pattern differences are due to accessions with least environmental factor.

Drought Effect on Growth and Development
There were significant differences among the accessions in respect to canopy temperature at 42 DAP.The highest mean canopy temperature was found in W-119 with an average of 27.24˚C while the lowest of 20.62˚C was recorded in 2003-24-2 (Table 3).These figures are quite low and attest to the drought tolerant capabilities of these accessions.A lower canopy temperature in a drought stressed plant indicate a better capacity for taking up soil moisture and for maintaining better plant water status [28].[29] used canopy temperature of drought stressed wheat accessions to characterize yield stability under various moisture conditions.A positive correlation was found between a DSI and canopy temperature in a stressed environment.Infrared canopy temperature provides an efficient method for rapid non-destructive monitoring of plant response to water stress.The average stem length of 38 cm which is the longest was recorded in Zapallo while the shortest was found in 2004-5-2 with a length of 28.6 cm.[30] observed that the reduction in stem length (relative to the control) of 15 accessions exposed to drought stress varied considerably from 16.1% to 46.0%.Internode diameter was reduced by 12% to 50% across the accessions.Ejumula had the highest number of leaves with a mean value of 37.4 while the least mean value was found in both Phala and 2004-16-1.
The severity of the stress condition became intense at 84 DAP (Table 4).There were also little or no differences in the responses of the accessions to the canopy temperature at that stage and the lowest temperature of 13.5˚C was recorded in 2004-9-2 while the highest was recorded in Ndou (17.8˚C).There was however a significant difference among the accessions in the stem lengths.The longest length of 40 cm was recorded in both Tacna and Ndou while the lowest length was found in 2004-5-2.The highest mean values for number of leaves were recorded in Zapallo (40.00), the lowest value of 22 was found in 2004-5-2.According to [31], drought stress significantly decreases Relative Water Content (RWC) and has a strong effect on photosynthetic rate and this also leads to increase in leaf and canopy temperature.For 42 DAP the highest root dry matter percentage was recorded in W-119 with a value of 23.09%, Ejumula followed with 22.94% (Table 3).The rest of the accessions did have a dry matter percentage ranging between 15 and 20 percent.There was a significant difference  among the accessions in the percentage shoot dry matter, the highest was recorded in 1999-3-1 having 25.63% and it was closely followed by W-119 and 2004-9-2 with 15.75%.The lowest percentage shoot dry matter of 14.17% was recorded in 2004-5-2.There was a general reduction in the root dry matter in all the accessions under water stress conditions.[32] indicated that storage root drymass is correlated positively with vegetative growth.Similarly, [11] reported a reduction in root dry mass under stress conditions.[11] reported a reduction in root dry mass under water stress condition.The variation in dry matter content can also be dependent on various factors such as soil type, pest, diseases, cultivar and climate [33].Despite the severity of the stress, some accessions still showed good traits reflecting in their dry matter accumulation at 84 DAP.The highest root dry matter of 60.90% was recorded in Resisto and this was followed by 2004-5-2 (59.39%) and Ndou (56.85%) (Table 4).Most of the rest of the accessions had mean values ranging between 37% to 49% with the lowest of 31.08 recorded in Tacna.Equally, Resisto had the highest shoot dry matter and the lowest was found in Tacna.
Observations made at 120DAP indicated that the accessions were matured for harvesting as they all showed signs of stress with most of the plant already wilting (Table 5).There was no significant difference in the canopy temperature among the accessions.2004-9-2 has the lowest canopy temperature of 26.2˚C while the highest canopy temperature of 30˚C -33˚C were found in W-119, 2003-24-2, 2004-16-1 and 2004-5-2.There was a noticeable difference in the stem length among the accessions; the longest stem length of 50 cm was recorded in Resisto while the shortest length of 27.0 cm was found in W-119.The highest number of leaves was found in Zapallo (60.6) while the lowest number was recorded in 2004-16-1 (14.8).

Drought Effect on Root Yield
Water stress is a common phenomenon and it severely reduces yields of field crops grown under rain fed conditions [34].The marketable yield in the present study ranged from 0.58 t•ha −1 to 3.83 t•ha −1 (Table 6).The highest marketable yield was recorded in Zapallo (3.83 t•ha −1 ), Tacna (3.63 t•ha −1 ) and Ndou (3.12 t•ha −1 ).The low- est marketable yield of 0.58 t•ha −1 was found in 2003-24-2.The ANOVA showed that there were significant differences among the accessions.Also, the three accessions with the highest marketable yield mostly had the highest total yield.The highest total yield was recorded in Tacna (9.24 t•ha −1 ), Zapallo (6.16 t•ha −1 ) and 2004-9-2 (5.58 t•ha −1 ).These yield values were comparable to the average yield values of 5-10 t•ha −1 normally recorded for sweetpotato grown under subsistence farming under rain-fed conditions.The lowest yield was recorded in 2004-5-2 (1.69 t•ha −1 ).This result is also consistent with the observations of [30] [27] on yield of sweetpotato accessions subjected to prolonged restriction of water.

Drought Effect on Survival Rates
Accessions Tacna and 1999-3-1 had the highest survival rate (Table 6) of 81.66% followed by W-119 (77.50%)Zapallo (77.49%),Ejumula (75.83%) and Ndou (75.0%).This result was an indication that these accessions had a very good mechanism to tolerate water stress.The lowest was found in 2004-5-2 (53.33%).It is therefore evident that yield and growth parameters such as leaf growth, dry biomass, root dry weight and internodes length can be used to successfully determine drought tolerant accessions in sweetpotato.In addition, [27] found a reduction in photosynthetic rate, stomatal conductance and expansion of leaf canopy related to reduction in yield of sweetpotato.

Table 1 .
Sweetpotato accessions used for box screening and their characteristics.

Table 2 .
Result of plastic box screening experiment.

Table 6 .
Yield data collected at harvest at 120 DAP.