Evaluation of cowpea mini core accessions for resistance to flower bud thrips Megalurothrips sjostedti Trybom (Thysanoptera: Thripidae)

Abstract The flower bud thrips, Megalurothrips sjostedti Trybom (Thysanoptera: Thripidae), is an economically important pest of cowpea in sub‐Saharan Africa. Varietal resistance is the most preferred, environmentally friendly, cost‐effective and sustainable option for controlling this pest. The objective of this study was to identify sources of resistance to M. sjostedti among mini core accessions from the largest world cowpea germplasm collection maintained at the International Institute of Tropical Agriculture (IITA). The study was conducted during the 2015 and 2016 cropping seasons where 365 accessions were screened under field conditions. Each accession was rated visually for thrips damage score, flower abortion rate, number of pods per plant and number of thrips per flower. The resistance levels observed in genotypes TVu8631, TVu16368, TVu8671 and TVu7325 were similar to that of the resistant check “Sanzisabinli” (called Sanzi) during both seasons. In addition, 56 mini core genotypes showed moderate resistance to thrips damage. High heritability values were associated with thrips damage scores at 65 days after planting (0.60), percentage of effective peduncles (0.59), flower bud abortion rate (0.59), number of pods per plant (0.51) and number of peduncles with pods (0.5). The accessions identified with good levels of resistance to flower bud thrips will be used in cowpea breeding programs to develop improved resistant varieties.

Cowpea is an attractive host to many insect pests that reduce its grain yield and quality. Among them, the flower bud thrips, Megalurothrips sjostedti Trybom (Thysanoptera: Thripidae), causes the most serious damage during the crop's flowering stage (Jackai & Daoust, 1986). The insect lays eggs on cowpea flower buds and the nymphs/adults feed on the reproductive structures of the plant (Alabi, Odebiyi, Tamò, & Omoloye, 2011), resulting in drying out and browning of the stipules, flower bud abscission, flower discoloration, distortion or abortion (Jackai, Inang, & Nwobi, 1992;Kanteh, Ndoleh, Dimoh, & Luseni, 2013). Due to premature flower abortion, the peduncles of susceptible plants are stunted as no pods develop on them. Grain yield reduction due to the flower bud thrips ranges from 20% to 80% (Omo-Ikerodah, Fawole, & Fatokun, 2008) and could reach 100% under high infestation (Alabi et al., 2011;Jackai & Daoust, 1986;Singh & Allen, 1980). Insecticide application is the predominant means of controlling this pest on cowpea. However, several alternative control measures including cultural practices (Ekesi, Maniania, & Ampong-Nyarko, 1999), biological control (Ekesi & Maniania, 2000;Mfuti et al., 2017;Tamò, Ekesi, Maniania, & Cherry, 2003) and the use of bio-pesticides such as neem extract (Badii, Nuamah, Braimah, & Awuku, 2016) have been explored to control this insect. Host plant resistance appears to be the most economical and environmentally friendly way to reduce thrips damage to cowpea. Unfortunately, most of the cowpea varieties grown in West Africa are highly susceptible to M. sjostedti. Only a few cowpea genotypes have been reported to show low levels of resistance to the pest (Omo-Ikerodah et al., 2008). Hence, a systematic evaluation of the currently available germplasm accessions could lead to the identification of more lines with higher levels of resistance to this pest. This study was conducted to evaluate the cowpea mini core accessions from the world's largest collection maintained at the Genetic Resources Center of IITA for new sources of resistance/tolerance to flower bud thrips. The genomic tools that are being generated in cowpea (Muñoz-Amatriaín et al., 2011) would further enhance the chances of successfully pyramiding these novel resistance genes in suitable varieties. Genotypes identified as thrips resistant will be used for the development of improved cowpea varieties, which would minimize the need for insecticide application by farmers when also ensuring increased grain yield.

| Study site
The study was conducted at Fashola, Latitude 7.9°N and Longitude 3.7833°E located between Oyo and Iseyin towns at about 60 km from Ibadan (Oyo State, Nigeria). The average temperature of the location during the main cropping season ranges from 21.42°C (T min ) to 31.82°C (T max ). The annual average relative humidity is 73.78%, while the annual average rainfall is about 1,173 mm. The soil type is Silt-clay. The location has been identified as a hotspot for flower bud thrips from previous field experiments conducted by IITA Cowpea Breeding Unit.

| Cowpea lines
The trials were carried out during 2015 and 2016 main cropping seasons (August-November) when cowpea is mostly grown by farmers in the area. A total of 369 cowpea germplasm accessions including 365 from the IITA mini core collection, one wild relative (TVNu699) and one Nigerian landrace (NGT65A) were screened under field conditions. A cowpea landrace from Ghana "Sanzisabinli" (called in short Sanzi) was used as resistant check because of its consistently low levels of damage scores and low population of thrips in flowers (Abudulai et al., 2006;Alabi et al., 2011;Sobda et al., 2018) while Vita 7 was used as susceptible check because of its well-known high damage scores and high thrips populations in flowers (Abudulai et al., 2006;Omo-Ikerodah et al., 2008).

| Planting and experimental design
During both years, seeds of the susceptible check, Vita 7, were sown in single row along the four sides of each replicate two weeks prior to sowing of test lines to help build up the population of the flower bud thrips in the field. No chemical application was done in the experimental field during the pre-flowering and flowering stages when cowpea is most susceptible to thrips. However, at podding stage, a synthetic pyrethroid insecticide, "Cyper-Diforce EC," composed of cypermethrin (30 g/L) and dimethoate (250 g/L) as active ingredi-

| Definition of the resistance status
Based on the damage score ratings (Jackai & Singh, 1988), accessions with scores less than five (i.e., 1-4) were considered resistant while those with scores of five and seven to nine indicated moderate level of resistance and susceptible, respectively.

| Statistical analysis
Data collected from the field and laboratory were subjected to analysis of variance using SAS 9.4 to determine if there were significant differences among the cowpea genotypes.

| Thrips damage scores and population
There were significant differences in thrips damage scores among the cowpea lines tested at 45, 55 and 65 DAP during both years (

| Resistance categories
Apart from the four lines that showed good levels of resistance to thrips, fifty-seven accessions made up of 56 from the mini core collection and one wild relative were regarded as moderately resistant. They had moderate damage scores of between 4.8 and 6.4 and relatively high number of pods (three to five pods/plants) except the wild relative, TVNu699 that had an average of 1.6 pods per plant.
Genotypes TVu8877 and TVu16521 despite the low damage scores were placed in this group because of their relatively low pod number per plant (Table 1).

| Number of peduncles with pods and flower abortion rate
Genotypes that had high number of peduncles with pods were characterized by low flower abortion rates. Among the genotypes considered as resistant, the number of peduncles with pods varied from 2.8 (Sanzi) to 5.4 (TVu9357) and the flower abortion rate ranged from 21.7% (TVu7739) to 55.1% (TVu8779). The susceptible line Vita 7 recorded 1.5% and 82.2% for number of peduncles with pods and abortion rate, respectively (Figure 1).

| Heritability of measured parameters
Following data analysis using the software Breeding View, it was possible to estimate heritability values for the different measured traits. High levels of heritability were associated with parameters such as damage scores at 65 DAP, pod number per plant, peduncles with pods, flower bud abortion rate and percentage of effective peduncles. The heritability values were comparatively low for damage scores recorded at 45 and 55 DAP and for thrips population in flowers (Table 2).

| Relationships between damage scores and other traits
The mean values for thrips damage scores ( Table 3).

| D ISCUSS I ON
The observed linear progression in thrips damage scores from 45 to 65 DAP in this study, can be attributed to a steady increase in the population density of the insect in the field from flower bud initiation to when the final scores were recorded as earlier suggested by . Flower bud thrips (M. sjostedti) infests cowpea plants from pre-flowering stage, and because of the insect's short developmental cycle of about 19 days (Salifu, 1992), it is able to produce up to four generations between flower bud initiation and flowering stages of the crop. The higher insect population at 65 DAP resulted in an increased pressure of the insect thereby causing more damage to the plant.
Symptoms of thrips attack on cowpea plants are well known. These include browning of stipules and flower buds, several stunted peduncles with no pods as well as flower bud abscission (Jackai & Singh, 1988). Severely infested plants appear diseased and produce only a very limited number of flowers that reach anthesis. This is because most of the flower buds drop prematurely (Singh & Allen, 1979).   study. We did not find any significant difference in thrips populations in flowers among the test entries; however, we observed that the damage recorded on the susceptible lines were higher than that on resistant ones. As shown in Table 2, there is a relationship between insect population in flowers and damages scores. This result corroborates that of Salifu, Hodgson, and Singh (1988), who found heavy infestation of thrips on susceptible lines at the flower bud stage leading to complete abortion of the flowers.
In this study, 57 accessions were identified as moderately resistant to M. sjostedti based on their moderate damage scores (which ranged from 4.5 to 6.3) and their average pod number per plant (three to five pods). Jackai and Singh (1988) following an earlier study had suggested that cowpea lines with damage scores of between five and six should be considered as moderately resistant to thrips. The mechanism of resistance operating in these genotypes could be a tolerance because they were able to produce some pods despite the fairly high levels of thrips damage observed on them in the field.
In this study, the resistant and moderately resistant genotypes displayed some characters such as high pod load and effective peduncles. These parameters are yield-related traits resulting from low flower abortion in thrips resistant/tolerant cultivars. According to Togola et al. (2017) The heritability values for some of the measured traits such as the damage scores at 65 DAP, the number of pods per plant, the number of peduncles with pods and the percentage of effective peduncles were high (>50%), implying that these traits are heritable and environmental effects on them are relatively low. Breeders can therefore make progress while developing improved cowpea varieties with resistance to flower thrips by using these traits as selection criteria. This is particularly so, when selection is based on low damage scores, low flower abortion rate and high number of effective peduncles. The results of this study show the superiority of the four resistant genotypes which can be as good as Sanzi, an already identified source of thrips resistance in cowpea.
These four resistant genotypes should be further evaluated to ascertain the types of genes coding for their resistance. Data from the genetic studies will enable breeders to effectively harness the resistance gene(s) for the development of improved cowpea varieties that are resistant to flower bud thrips.

| CON CLUS ION
The need for the identification of sources of resistance to M. sjostedti in cowpea has remained a front burner in cowpea breeding programs in SSA. With the identification of several promising lines characterized by good levels of resistance to flower bud thrips, development of improved cowpea with resistance to this pest appears feasible in the foreseeable future. It is conceivable that different gene loci could be responsible for the resistance detected among some of the 4 lines in this study. It will therefore be necessary to determine if allelism exists among some or all of the resistant lines. From this study, those lines not showing allelism could be crossed for the purpose of pyramiding the genes present in the development of thrips resistant varieties. This will ensure a robust and more durable resistance in improved varieties containing the pyramided genes. The high number of genes will serve as buffer in controlling thrips damage in the new improved varieties.

ACK N OWLED G EM ENTS
The authors acknowledge the funding for this study by Bill and