Breeding for bean fly resistance in common bean (Phaseolus vulgaris L.): a review

ABSTRACT

The bean fly (Ophiomya spp) is the most important yield limiting insect pest of common bean in Africa. The insect pest can cause complete crop loss affecting bean production and productivity under epidemic conditions. Effective control of bean fly is essential for sustainable bean production Africa. The overall progress, opportunities and challenges of the bean fly control strategies. The biology and ecology of bean fly and the economic importance of the insect pest is presented as well as the existing controlling strategies, with an emphasis on the breeding on breeding strategies used, research progress achieved challenges and opportunities. In conclusion, significant research progress have been made in breeding for bean fly resistance evident by identification of breeding lines, understanding resistance mechanism and development of breeding strategies. However, there is a need for further research to validate the available information and also explore new breeding methods such molecular breeding which has not been explored at present. Such studies will accelerate breeding for bean fly resistance.

KEYWORDS:

• Bean fly
• common bean
• insect pest
• Ophiomya spp
• Phaseolus vulgaris
• resistance breeding

Introduction

The bean fly (Ophiomyia spp.), also known as the bean stem maggot, is the most important and destructive insect pests of common bean. The pest is widely reported in Africa, Asia and Australia (Kornegay and Cardona 1991). Under epidemic proportions the pest can cause total crop loss. Hence it is a key yield limiting factor of bean production and productivity (Kiptoo et al. 2016). Globally, the bean fly has received little research attention compared with other common bean production constraints such as bean diseases. Most research on the taxonomy, biology and ecology of insect was conducted during the early 1930s and 1990s (Van der Goot 1930; Frick 1952; Greathead 1969; Abate, 1990a). Several recent studies have focussed on identifying resistant genotypes, understanding the inheritance mechanism of resistance and on the best agronomic control practices (Letourneau 1995; Mushi and Slumpa 1996; Rahaman and Prodhan 2007; Ojwang et al. 2010; Ssekandi et al. 2016). Attempts have also been made to develop methods of rearing the insect for artificial screening of common bean lines under greenhouse conditions (Kiptoo et al. 2016). However, there is no standard protocol for artificial rearing, inoculation and assessment of the bean fly. Despite advances in the application of molecular plant breeding tools for some major insect pest resistance traits, genetic markers have yet to be developed to assist in the selection of bean fly resistance (Miklas et al. 2006). Hence, the objective of this review is to provide progress, opportunities and challenges in the development of effective bean fly control strategies. The paper outlines the biology and ecology of bean fly and the economic importance of insect pest with an emphasis on breeding strategies, selection methods, research progress achieved and challenges and opportunities.

Taxonomy and identification of the bean fly

The bean fly belongs to the Dipteran family Agromyzidae. The family consists of the following species: Ophiomyia phaseoli Tyron (O. phaseoli), O. spencerella Greathead (O. spencerella), O. centrosematis de Meij (O. centrosematis), Melanagromyz sojae Zehntner (M. sojae), M. phaseolivora Spencer (M. phaseolivora) (Allen and Smithson 1986). The first three species are the most destructive, whereas the last two are either minor or occasional pests of common bean (Talekar 1992). As such, very little published information exists on the biology, damage and control of the latter three species (Talekar 1992).

All adults of the bean fly species have a similar appearance. Morphologically, the insect is a tiny black fly that is difficult to observe in a bean field (Figure 1). The key descriptors are the location of larval feeding sites in the host plants, shape and size of the insects’ anterior and posterior larval and pupal spiracles (Talekar and Chen 1985). The larva of O. phaseoli has small anterior spiracles that consist of a circle of six minute bulbs and their posterior spiracles are closely adjoined on a conical projection with about 10 bulbs. The puparium is pale yellow or brown, and 3 mm in length. The adult bean fly mainly feeds and pupates in the main stem between the outer epidermis and the cortex. The morphological features of posterior spiracles of O. spencerella are identical with those of O. phaseoli. The only external character that distinguishes O. spencerella from O. phaseoli is the shiny black pupae of the former, while the latter has pale yellow to brown pupae. The shiny black feature of the pupae of O. spencerella can be seen underneath the epidermis where the larvae feeds and pupates. Ophiomyia centrosematis larva has anterior spiracles on long black stalks and posterior spiracles on short stalks, each with three bulbs, while the puparium is yellow or red – orange. Feeding and pupation mainly take place on the main stem between the outer epidermis and the cortex (Greathead 1969). The major species of the bean fly are described below.

Figure 1. A bean fly insect on a bean plant at Chitedze Research farm in Malawi.

Ophiomyia phaseoli Tryon

O. phaseoli is the most predominant and destructive of the three bean fly species found across the continents of Africa, Asia and Australia (Abul-Nasr and Assem 1968; Talekar 1992; Kayitare 1993). It attacks a wide range of legumes including the common bean (Phaseolus vugalis L.), soybean (Glycine max L.) and cowpea (Vigna unguiculata [L.] Walp.) (Abul-Nasr and Assem 1968). Sunflower (Carthamus tintorius L.) and black night shade (Solanum nigrum L.) are among the hosts of the devastating insect pest (Kayitare 1993). The adult insect is usually between 1.5–2 mm in length with a wing span of approximately 3 mm (Karel and Autrique 1989). Its life cycle has four stages namely the egg, larva, pupal and adult fly stages. The fly is more active in shady sunny environments than in hot sunny areas. It is susceptible to dehydration, hence it has a shorter life cycle (3 weeks) in hot weather when compared to cool weather (12 weeks) (Davis 1969; Kornegay and Cardona 1991). The adult female lays oval, milky white, opaque or translucent eggs via the ovipositor 2–3 days after copulation. Oviposition occurs during day light hours on either the upper or lower surfaces of young bean leaves. Oviposition on the lower surface of the leaf epidermis occurs during the rainy season (Davis 1969). The ovipositor’s labellum is used to identify a suitable host plant where the female insect creates punctures on the epidermis. Approximately one of the 10 punctures contains eggs, while the rest serve as feeding punctures (Manohar and Balasubramanian 1980). A single bean fly can lay between 100 and 300 eggs in two weeks. The incubation period is 2–4 days and the duration of the larval and pupal stages ranges from 8–10 and 9–10 days respectively (Kayitare 1993). The pupation period varies from one geographical location to the other. In the tropical lowlands the pupation period is shorter (7–13 days) than in the tropical highlands (13–20 days) (Abate 1990a). The larva creates feeding tunnels and pupation takes place in these tunnels. The adult fly emerges in the morning hours and is initially light brown before it turns black. Premating period in adults last for three days with variable copulation periods of between 40 min and 3 h (Singh 1982). The insect has a unique feeding habit such that the females feed on the exudate from the wound and the males feed on the remains, after the females have left the plant (Kayitare 1993).

Ophiomyia spencerella Greathead

Ophiomyia spencerella was first observed as an insect pest of P. vulgaris in 1968 (Greathead 1969). It is indigenous to Africa and more important in East Africa (Greathead 1969). Apart from common beans, the insect pest also attacks other members of the Fabacea family. Although in small numbers, it has also been found in rice bean (Vigna umbellate Thunb), lima bean (Phaseolus lunataus L.), black gram (Vigna mungo [L.] Hepper) and cowpea (V. unguiculata [L.] Walp.) (Greathead 1969). The females O. spencerella mainly oviposit the eggs into the hypocotyl at ground level when the seedling is two to three days old. Few eggs are deposited in young stems above the cotyledons and rarely in the leaves. The larvae feed near the ground or on the taproot moving back to the ground to pupate. Adults emerge through thin transparent layers that are made through the epidermis (Kayitare 1993). At 21°C, the eggs develop into adults in 28–37 days (Greathead 1969).

Ophiomyia centrosematis de Meij

Ophiomyia centrosematis is widely distributed in East Africa, Australia and Tropical Asia (Spencer 1973). Apart from common bean, this fly has a wide range of host species viz. calopo (Calopogonium muconoides Desv), butterfly pea (Centrosema pubescens Benth) and cowpea (V. unguiculata L) (Greathead 1969). The female adult oviposits an average of 63 eggs in the hypocotyl just beneath the epidermis (Talekar and Chen 1985). The larvae feeds just beneath the epidermis of the stem and tap root, but do not tap into the pith of the root or stem. The larval stage lasts up to 11 days and during the larval period the larva undergoes three instars. Pupation generally occurs close to the soil surface where the translucent red to yellow brown coloured pupa, is lodged beneath the epidermis. The anterior spiracles of pupae pierce the dry epidermis and form a semi-transparent layer to facilitate adult emergence (Greathead 1969; Kayitare 1993). After 11 days of pupation the adult emerges. Mating starts within 2–3 days of the adult stage, females start laying eggs in 3–4 days after copulation and oviposition last up to 18 days. Adults make oviposition and feeding holes in the hypocotyl and feed on the sap oozing from such holes. On average there are 3–4 generations per cropping season (Talekar 1992)

Geographical distribution and seasonal occurrence of the bean fly

The occurence of bean fly varies from location to location depending on the climate and host plant availabilityand its incidence has been reported in several countries (Table 1). Serious damage occurs during the dry season. A minimum temperature of 15°C, a maximum of 32.5°C, relative humidity of 65–75% and low rainfall are favourable conditions for a bean fly epidemics. O. phaseoli development in the field is inhibited by high precipitation. On the other hand, high temperatures are reported to increase the incidence of insects (Talekar and Chen 1985). In soybean, O. phaseoli and O. centrosematis attack the crop throughout the year, but the severe infestation occurs during dry seasons (Talekar and Chen 1985). Precipitation negatively affects the feeding and oviposition, resulting in low infestation. Low winter temperatures result in inactive pupae, while extremely hot temperatures hinder egg hatching. Abul-Nasr and Assem (1968) reported low infestations on crops that were grown in the winter months and high infestations on crops that were grown during hot periods. Information on the incidence of bean fly pestilences in rain fed agriculture could be used in monitoring the insect pest. In Kenya, late planting was associated with a high incidence of bean fly when compared with early planting (Songa 1999). Although the distribution of the species overlap, they seem to have different environmental preferences. Ophiomyia spencerella prefers medium to high altitude areas receiving high rainfall, while O. centrosematis prefers low altitude and O. phaseoli is distributed in all environments (Songa 1999).

Table 1. Countries where bean fly has been reported.

Country	Reference
Australia	Morgan (1938)
Botswana	Greathead (1969)
Burundi	Autrique (1989)
China	Wang and Gai (2001)
Egypt	Abul-Nasr and Assem (1968)
Ethiopia	Abate et al. (1995)
India	Agarwal and Pandey (1961)
Indonesia	Talekar and Chen (1985)
Japan	Yasuda (1979)
Kenya	Kiptoo et al. (2016) and Ojwang et al. (2010)
Malawi	Letourneau (1994) and Kapeya et al. (2005)
Mozambique	Davies (1998)
South Korea	Kwon et al. (1980)
Taiwan	Talekar and Shyan Chen (1983)
Tanzania	Greathead (1969)
Thailand	Van der Goot (1930)
Uganda	Ssekandi et al. (2016)
Vietnam	Talekar and Chen (1985)
Zambia	Karel (1985)
Zimbabwe	Karel (1985)

Damages caused by the bean fly

The primary indications of bean fly damage are dead seedlings or weak and stunted plants, stems become dry or swollen and split (Allen et al. 1996). Another common symptom of bean fly infestation is yellowing of the leaves at an early plant stage. Seriously infested plants are characterised by premature leaf drop and death (Karel and Autrique 1989). The major damage comes from feeding of the larvae inside the stem. Damage caused by adult feeding, although visible, is insignificant. The third instar larvae destroy the medullary tissue of the stem at the ground level. Plant damage is more pronounced in dry conditions than wet conditions. Later in the life of the plant, the larvae cause little damage (Karel and Autrique 1989). All the species of the bean fly are internal stem borers and feed beneath the stem epidermis where pupation takes place. Out of the three species, O. centrosematis cause the least damage (Karel and Autrique 1989; Kornegay and Cardona 1991).

The most visible adult damage in common bean is more pronounced at unifoliate leaf stage. The unifoliate leaves show large number of feeding and ovipositional punctures on the upper side with corresponding light yellow spots, especially on the basal portion of the leaf. Under severe attack, the unifoliate leaves become prematurely yellow and usually drop off (Kornegay and Cardona 1991). The larval damage which has high impact on physiology of plant or seed yield starts in the leaf lamina. However, more pronounced damage occurs when the larvae start feeding underneath of the stem epidermis. Under severe conditions, the stem becomes swollen with cracked skin and cankerous surface. Total destruction of the cortex tissue weakens the stem, resulting in lodging of the plant. The lodged plants do not recover and this results into considerable yield loss. In many cases this damage results in plant mortality within 3–4 weeks after germination. If part of the cortex tissue remains intact, the plant continues to grow and develops a new root system above the point of injury by forming adventitious roots. In wet weather the lowest adventitious roots can reach considerable lengths and can compensate for the loss of a large part of the root system (Talekar 1992).

Yield loss due to bean fly in common bean and legumes is poorly documented, although some researchers have reported that it ranges from 10% to 100% (Table 2) (Kapeya et al. 2005; Ojwang et al. 2010). The extent of yield loss depends on the bean fly species, levels of infestation, time of infestation, susceptibility of the bean cultivars as well as environmental factors. O. phaseoli causes significant yield loss than any other bean fly species followed by O. spencerella (Abul-Nasr and Assem 1968; Talekar 1992; Kayitare 1993). Yield loss and damage are more pronounced when plants are attacked at the early seedling stage when compared to an attack at a later growth stage (Karel and Maerere 1985). In Malawi, Kapeya et al. (2005) observed 90% mortality rate of bean plants attacked at the seedlings stage. Yield loss of 33% was recorded, when the crop was attacked at the late seedling stage in Kenya (Karel and Maerere 1985).

Table 2. Reported yield loss due to bean fly infestation in legumes.

Crop	Yield loss (%)	Country	Reference
Common bean	60–80	Mozambique	Davies (1998)
	10–100	Malawi	Kapeya et al. (2005)
	30–100	Kenya	Ojwang et al. (2010)
	33–100	Tanzania	Karel (1985)
	8–37	Ethiopia	Abate (1990b)
Soybean	35	Taiwan	Talekar (1989)

Control strategies of the bean fly

Various control strategies are suggested to minimise losses incurred by the bean fly infestation in legume crops. The control options are briefly described below.

Biological control

Biological control involves the use of parasites of Ophiomyia species. The parasitic insect lays eggs in the late larval instars and the adult emerges from the pupae. Opius phaseoli Fisher and O. importatus Fisher have been used as a biological control agent of bean fly. Of these species, O. phaseoli Fisher has been reported to be the most effective agent in reducing the population of O. phaseoli (Talekar 1992). The efficiency of these parasites has been linked to a location effect (Talekar 1992). For instance, in Kau and Maui island of Hawaii 100% and 25–83% parasitism of O. phaseoli was recorded, respectively (Davis 1972).

Cultural control

Cultural control practices include ridging, increased plant density, fertiliser application to encourage vigorous plant growth, intercropping, adjustment of planting dates, crop rotation and mulching to cover the planting area (Karel and Autrique 1989; Letourneau 1994; Songa 1999; Kapeya et al. 2005). Intercropping of maize and common bean reduces bean fly infestation, especially when maize is planted earlier than common bean (Talekar 1992). The varietal mixture is another cultural control strategy that farmers have recently adopted Ssekandi et al. (2016).

Chemical control

Several insecticides such as dimethoate, cypermethrin and pyrethrum are used in controlling the bean fly population. The efficacy of insecticides depends on time of application, mode of application and type of insecticides. An early application of suitable chemicals mostly gives good results (Talekar 1992; Kapeya et al. 2005; Rahaman and Prodhan 2007; Ssekandi et al. 2016). Seedlings from treated seed usually have insecticide residues, which protect the crop from early damage by bean fly. Systemic insecticides are quickly absorbed in the plant tissue, where larvae, pupae and eggs could be easily killed. Despite insecticides being recommended as an effective way of controlling bean fly, they are unaffordable to resource poor farmers and are unfriendly to the environment (Kapeya et al. 2005; Ssekandi et al. 2016).

Host plant resistance

The use of resistant cultivars, together with other available control methods, is the most effective way of controlling the bean fly (Karel and Autrique 1989). Host plant resistance to insects (HPRI) is an effective, economic and environmentally friendly method of pest control (Sharma and Ortiz 2002; Richardson et al. 2006). However, there are few market class insect-resistant cultivars when compared to disease-resistant cultivars despite the importance of HPRI (Miklas et al. 2006). The major contributing factors to slow progress in developing insect-resistant cultivars include (1) few studies have been conducted on the genetics of the insect pest resistance, (2) difficulties associated with ensuring adequate insect infestation for resistance screening studies, (3) high efficiency of insecticides, (4) lack of interdisciplinary team work between breeders and entomologists, (5) poorly established agricultural research in the developing world and (6) rapid evolution of the insect pests (Kornegay and Cardona 1991; Sharma and Ortiz 2002; Miklas et al. 2006). Insect-resistant cultivars are able to inhibit oviposition and feeding, reduce insect survival and development, and tolerate or recover from insect damage (Smith 1989; Sharma and Ortiz 2002).

Plants exhibit three forms of resistance to insect pests namely; antibiosis, antixenosis and tolerance. The antibiosis type of resistance reduces the survival and reproduction of the insect or prolongs the time between the generations. Antixenotic resistance reduces the rate of both initial and successive insect population build-up. Tolerance is when plants have the ability to grow and yield fairly well despite being attacked by the pest (Cardona and Kornegay 1999; Sharma and Ortiz 2002). The most common resistance mechanism to bean fly is tolerance to stem damage (Kornegay and Cardona 1991).

Bean fly resistance is a polygenic and durable resistance which is easily attained when genes are pyramided, resulting in the simultaneous expression or interaction of more than one gene (Joshi and Nayak 2010). Pyramiding multiple genes of insect resistance is a promising approach. However, very few attempts have been made in the past (Miklas et al. 2006). Bueno et al. (1999) incorporated multiple resistance genes into common bean cultivars for Empoasca, Apion and Zabrotes. Similarly, introgression of bean fly resistance genes into market class cultivars has been reported in Tanzania (Hillocks et al. 2006). Gene pyramiding for bean fly resistance has not yet been reported despite the identification of several common bean resistant genotypes.

Components of bean fly resistance

One major challenge in breeding for bean fly resistance is the lack of systematic screening procedures that exert uniform infestation of the genotypes (Hillocks et al. 2006). This has been a challenge in the progress in breeding for resistance to bean fly (Sharma and Ortiz 2002). Most plants, including legumes, rely on a number of defence mechanisms for protection against insects (Edwards and Singh 2006). These set of defence mechanisms are either biochemical compounds or morphological features. Significant correlations between resistance to bean fly and morphological features such as internode length, leaf hairiness and stem diameter in common bean and other related species have been observed (Rogers 1980; Abate 1990a).

In soybean, varieties resistant to bean fly were found to have a 10 times higher concentration of phenolic compounds compared to susceptible varieties. Phenolic compounds serve as toxicants that inhibit the growth of bean fly (Kayitare 1993). Phenolic compounds, such as tannins, make the plant unpalatable. Plant architecture has also been associated with resistance to bean fly. Resistant cultivars tend to have short internodes in early stages of the plant growth. Short internodes increase the toughness of stem tissues such that the insect pest has difficulties in chewing the tissue (Abate 1990b; Kayitare 1993). Bean fly resistance is correlated to the cross section of plant pith. It was observed that resistant cultivars in soybean had small pith, almost 1.5–2 times smaller than that of susceptible ones (Chiang and Norris 1984). A narrow pith with tightly enclosed lignified gelatinous xylem fibres restricts the movement of the larvae to pass through, or turn around to continue feeding until it pupates. Chiang and Norris (1982) reported a negative correlation between the toughness of the stem and the number of bean flies that damage mung bean. There is a relationship between stem colour and bean fly resistance (Chiang and Norris 1984). Cultivars with purple stem colour were found to be resistant in soybean (Chiang and Norris 1984). Purple pigment in the epidermis of soybean has anthocyanidin malvidin, which contains some phenols that contribute to resistance in purple stemmed soybean. Abate (1990a) observed antibiosis resistance in some bush bean accessions. The resistant accessions exhibited symptoms of bean fly damage, however, a high mortality rate of the bean fly was recorded in such accessions. Cultivars with high root density have also been reported to be tolerant to bean fly when compared to those with a low root density (Karel and Autrique 1989). Breeders could use these traits that are associated with bean fly resistance as tools for indirect selection in bean fly resistance programmes in the absence of reliable screening techniques.

Resistance breeding against the bean fly

Sources of genetic variation for bean fly resistance

The availability of good sources of resistance is a prerequisite for successful resistance breeding programmes. Searching for sources of resistance to bean fly dates back to the early 1940s (Moutia 1944). The exploration of resistant sources has not been limited to P. vulgaris but across Phaseolus species. At present sources of resistance have been identified both from primary and secondary gene pools (Kornegay and Cardona 1991). Sources of resistance from primary gene pool are reported to possess a high level of resistance (Cardona and Kornegay 1999). These identified sources (Table 3) have been transferred into the cultivated common bean through interspecific hybridisation (Karel and Autrique 1989; Kornegay and Cardona 1991; Ojwang’ et al. 2011a). In Africa, common bean landraces have been important sources of resistance to bean fly. In Kenya, (Ojwang’ et al. 2011b) crossed GBK047821(landrace)/KatB69(market class cultivar) and GBK04785 (land race)/Kat x 69(Market class cultivar) in attempt to develop bean fly resistant market class cultivar.

Table 3. Summary of some bean genotypes identfied as sources of resistance to bean fly in legumes.

Cultivar/line	Legume species	References
G5733, G2072	Phaseolus vulgaris L.	Abate (1990a) and Abate et al. (1995)
ZPV292, G5773, A55, G2005	Phaseolus vulgaris L.	Mushi and Slumpa (1996)
G35023, G35075	Phaseolus coccineous L.	Kornegay and Cardona (1991)
A429, TMO	Phaseolus vulgaris L.	Abate et al. (1995)
Mlama 49, Mlama 127, G22501	Phaseolus vulgaris L.	Hillocks et al. (2006)
GBK 047810, GBK 047866 GBK 047821, GBK 036488	Phaseolus vulgaris L.	Ojwang et al. (2010)
G21212,CIM 9314-36, Ikinimba	Phaseolus vulgaris L.	Mushi and Slumpa (1996)
G2472,EMP81, G3844, BAT 16	Phaseolus vulgaris L.	Mushi and Slumpa (1996) and Ojwang et al. (2010)
KK 8, Tasha, Chelalang	Phaseolus vulgaris L.	Kiptoo et al. (2016)

Evaluation for bean fly resistance

Accurately, efficient techniques that identify plants with insect resistance are essential to all insect resistance plant breeding programmes (Sharma et al., 2003). Plant resistance to an insect is measured through the exposure of a plant or plant parts to the insect, and is generally evaluated as the percentage of damage to the plant foliage, reduction in stand or grain yield, and general vigour of the plant (Sharma et al., 2003). Evaluation techniques based on measurement of insect damage to plants vary with crop species, pest insect and site (e.g. under laboratory, greenhouse and field conditions). For successful insect pest resistance screening programme, the selected screening method should give distinctly different reactions for plants of susceptible, moderate and resistant cultivars. Several studies of screening for resistance to bean fly have relied mainly on field screening followed by the greenhouse screening (Distabanjong and Srinives 1985; Ojwang et al. 2010; Kiptoo et al. 2016).

Screening for bean fly resistance in the field has mainly relied on natural infestation (Oree and Slump 1990; Ojwang et al. 2010; Kiptoo et al. 2016). However, due to the variation in insect populations over time and space, it is difficult to identify reliable and stable sources of resistance under natural infestation (Sharma and Ortiz 2002). To improve the efficiency of field screening, several techniques have been exploited including (1) planting in hotspots, (2) late planting and (3) planting susceptible cultivars as infector rows. These techniques subject the crop to uniform insect pressure at the most susceptible stage of crop development (Sharma and Ortiz 2002; Hillocks et al. 2006). Though not well exploited in bean fly resistance breeding, greenhouse screening permits greater control in selecting resistant plants and allow mass screening of seedlings thereby saving time. While screening for bean fly resistance in Kenya, Ojwang (2010) and Kiptoo et al. (2016) exploited the greenhouse screening technique using the no choice test.

Damage rating to bean fly infestation

Host reaction scales are generally developed in germplasm screening programmes to accurately describe insect damage levels (Smith et al. 1993). An ideal scale is the one that establishes a distinction among the small differences in plant damage by clearly defining resistant, moderately resistant, and susceptible genotypes. Scales that are fast and easy to use are recommended considering a large size of plant population managed by breeding programmes (Smith et al. 1993). The use of visual damage descriptors is one way that makes a damage rating scale user friendly. Several damage rating scales for bean fly damage evaluation were proposed in 1987 (Tables 4 and 5) (Kornegay and Cardona 1991). These rating scales have been used in several bean fly evaluation studies (Ojwang et al. 2010, 2011b; Kiptoo et al. 2016). The 1–9 severity scale has been exploited most often due to its flexibility. In general scores of 1–3 indicate that the genotypes are resistant, 5 moderate resistant, while 7–9 to susceptible. The severity rating scales are used in different growth stages of the bean crop. Pupae and larva counts are also used in assessing bean germplasm for bean fly. Cultivars with high pupa and larvae numbers are reported to be susceptible. However, some genotypes may have high pupa counts but tolerant to bean fly (Abate 1990b).

Table 4. Severity score for bean fly infestation at flowering stage on plot basis.

Rating	Category	Symptoms
1	Resistant	Infested plants are vigorous as uninfested plants.
		The bean fly apparently causes no considerable damage.
3		Infested plants with slight growth delay.
5	Moderately resistant	Infested with considerable growth delay.
7	Susceptible	Infested plants with severe growth delay.
9		Infested plants dead or almost dead.

Source with modification; Kornegay and Cardona (1991).

Table 5. Damage score for 10 individual plants per plot (Including shoot, stem and roots).

Rating	Category	Symptoms
1	Resistant	No damage; Stems, roots and shoots of infested plants look as the ones of uninfested ones, i.e. no cracks (stems and roots).
3		Infested plant stems, roots and shoots slightly damaged, i.e. some few cracks (stem and roots).
5	Moderately resistant	Infested plant stems, roots and shoots with considerable cracks, i.e. moderately damaged causing considerable growth delay.
7	Susceptible	Stems of infested plants with huge cracks causing stunted growth.
9		Stems with very severe cracks and destroyed root system leading to plant death.

Source with modification; Kornegay and Cardona (1991).

Genetics of bean fly resistance

The mode of inheritance of resistance to insects is not well studied when compared to disease resistance in common bean (Miklas et al. 2006). Lack of information has made breeding for resistance to the insect more complex, which has resulted in few cultivars being bred specifically for insect resistance when compared to those bred for disease resistance in common bean (Miklas et al., 2006). The mode of gene action conditioning inheritance of resistance or tolerance to bean fly indicated that it is polygenic with an involvement of both major and minor genes (Table 6) (Mushi and Slumpa 1996; Wang and Gai 2001; Ojwang et al. 2011a). Major genes effects have been found to be more significant than minor genes effects for bean fly inheritance (Wang and Gai 2001). Contrary to these findings, resistance to bean fly was reported to be conditioned by one major gene in soybean (Wei et al. 1989). Ojwang et al. (2011a) and Mushi and Slumpa (1996) reported that additive gene action was more important in resistance to bean fly than dominance gene action in common beans. The importance of additive gene action varied from parent to parent, justifying a need to test different parents (Mushi and Slumpa 1996). In mung bean additive, dominance and epistatic gene effects in the resistance to bean fly were observed (Distabanjong and Srinives 1985). Similar results were observed in soybean (Wang and Gai 2001). An understanding of the mode of gene action conditioning the expression of traits of interest is fundamental in breeding programmes because it helps in devising breeding strategies for crop improvement (Zalapa et al. 2006). Low to moderate heritability value, ranging from 0.22 to 0.45, has been reported for bean fly resistance in common bean (Ojwang’ et al. 2011b). In soybean, Wang and Gai (2001) observed heritability values of 0.20–0.80 for bean fly resistance.

Table 6. Summary of reported gene action conditioning resistance to bean fly.

Crop	Gene action	References
Common bean	Additive	Mushi and Slumpa (1996) and Ojwang et al. (2011b)
	Dominance	Mushi and Slumpa (1996) and Ojwang et al. (2011b)
	Epistasis	Ojwang et al. (2011b)
Mung bean	Additive	(Distabanjong and Srinives (1985))
	Dominance	Distabanjong and Srinives (1985)
	Epitasis	Distabanjong and Srinives (1985)
Soybean	Additive	Wang and Gai (2001)
	Dominance	Wei et al. (1989) and Wang and Gai (2001)
	Epistasis	Wang and Gai (2001)

Breeding strategies for bean fly resistance

As reviewed by Singh and Schwartz (2010b), principal factors determining strategies and methods used for resistance breeding to insect pest such as bean fly and other desirable traits include (1) the genetic distance between the cultivar to be improved and the resistant germplasm, (2) the direct and indirect screening methods available, (3) the genetics of resistance and (4) the number traits to be improved. Wide genetic diversity between the recipient and donor parent usually demand integration of two or three breeding approaches. The pedigree, mass selection and single seed descent breeding methods are reported to be ideal and effective methods for transferring major resistance alleles and quantitative traits loci (QTL) between cultivars and breeding lines within the market classes (Table 7) (Singh and Schwartz 2010b). Some form of backcrossing such as recurrent backcrossing, inbred backcrossing or congruity backcrossing (backcrossing alternately with either parents) becomes essential as the genetic distance between the cultivar under improvement and the resistant donor germplasm increases (Singh and Schwartz 2010b). In attempts to breed for resistance to bean fly and also to understand the genetics of bean fly resistance in mung bean and soybean, backcross breeding has been deployed in several studies (Distabanjong and Srinives 1985; Wang and Gai 2001). Ojwang et al. (2011b) used pedigree selection in advancing common bean progenies developed from crosses involving bean fly resistant parents in Kenya and market class beans. Attempts were made to improve resistance to bean fly through the interspecific crossing of P. coccinneus L and P. vulgaris L in some breeding programmes. Singh and Schwartz (2010b) indicated that breeding programmes involved in introgressing genes from other gene pools should use backcrossing or recurrent selection.

Table 7. Summary of some recommended selection methods for bean fly resistance improvement.

Selection methods	Reference
Pedigree	Singh and Schwartz (2010a) and Ojwang’ et al. (2011b)
Single seed descent	Mushi and Slumpa (1996)
Recurrent selection	Distabanjong and Srinives (1985), Singh and Schwartz (2010a)
Backcrossing	Distabanjong and Srinives (1985), Wang and Gai (2001) and (Singh and Schwartz, 2010a)

Application of genomic tools in breeding for bean fly resistance

In common bean, molecular breeding has been more commonly exploited in breeding for disease resistance when compared to insect resistance (Miklas et al. 2006). Few studies to map QTLs for insect resistance on the common bean genome have been conducted (Murray et al. 2004; Frei et al. 2005; Blair et al. 2006). Blair et al. (2006) developed nine molecular markers for Apion resistance and mapped them to loci on linkage groups B1, B8, B7 and B11 based on genetic analysis of a recombinant inbred line (RIL) population (Jamapa/J-11). Major QTLs conditioning resistance to thrips have been identified on linkage groups B2, B3, B6 and B8, using RILs derived from a BAT 881/G 21212 population (Frei et al. 2005). Genes conditioning resistance to bruchids (Zabrotes sub-fasciatus) and leafhoppers of genus Empoasca (Homopotera: Cicadellidae) have been tagged in independent studies by Osborn et al. (1986) and Murray et al. (2004), respectively. No genetic tagging studies have been conducted for bean fly resistance. Furthermore, there are no genetic markers developed for selection for bean fly resistance (Miklas et al. 2006). However, the success reported on other bean insect pests in terms molecular marker development is a positive indication that the same results may be achieved for bean fly resistance. Recently, a study to tag genes conditioning resistance to bean fly using RILs developed from the cross A55/G122 is underway (data not presented).

Conclusion and future breeding outlook

Significant research progress have been made in breeding for bean fly resistance evident by identification of breeding lines across the Phaseolus gene pools, understanding resistant mechanism and development of breeding strategies. Despite the reported progress, there is a need for identifying more breeding lines, validating reported breeding mechanisms and selection strategies. In addition, molecular breeding has not yet been explored in bean fly breeding to the extent that there are no markers of bean fly resistance. Hence, there is a need to develop genomic tools for screening bean fly resistance genes and marker-assisted breeding. Such studies will result in the identification of best breeding strategy, accelerated breeding and increased breeding efficiency for bean fly resistance.

Acknowledgments

The authors thank the Chitedze Agricultural Research Station for hosting of the first author.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Wilson Nkhata is a Ph.D. fellow in Plant Breeding at University of KwaZulu-Natal in South Africa.

Hussein Shimelis (Ph.D.) is a Professor of Plant Breeding and Deputy Director of the African Centre for Crop Improvement (ACCI) at the University of KwaZulu-Natal, South Africa.

Rob Melis (Ph.D.) is Professor of Plant Breeding with African Centre for Crop Improvement (ACCI) at the University of KwaZulu-Natal, South Africa.

Rowland Chirwa (Ph.D.) is a Plant Breeder and Coordinator of Southern African Bean Research Network (SARBN) at Centre for International Tropical Agriculture (CIAT), Malawi.

Tenyson Mzengeza (Ph.D.) is a Plant Breeder and Senior Deputy Director of Agricultural Research Services at Ministry of Agriculture Irrigation and Water Development, Malawi.

Additional information

Funding

The authors thank the Alliance for Green Revolution in Africa (AGRA) for financial support [grant number AGRAPASS30].