Biology and integrated pest management for the banana weevil

Integrated Pest Management Reviews 6: 79–155, 2001.© 2003 Kluwer Academic Publishers. Printed in the Netherlands.

Biology and integrated pest management for the banana weevilCosmopolites sordidus (Germar) (Coleoptera: Curculionidae)

Clifford S. Gold1, Jorge E. Pena2 & Eldad B. Karamura3

1IITA-ESARC, P.O. Box 7878, Kampala, Uganda2TREC, University of Florida, 18905 SW 280th Street, Homestead, Florida 33031-3314, U.S.A.3INIBAP/ESA, IPGRI, P.O. Box 24384, Kampala, Uganda

Key words: banana, banana weevil, Beauveria bassiana, biological control, Cosmopolites sordidus, culturalcontrol, host plant resistance, integrated pest management, neem, plantain

Abstract

The banana weevil Cosmopolites sordidus (Germar) is the most important insect pest of bananas and plantains(Musa spp.). The larvae bore in the corm, reducing nutrient uptake and weakening the stability of the plant. Attackin newly planted banana stands can lead to crop failure. In established fields, weevil damage can result in reducedbunch weights, mat die-out and shortened stand life. Damage and yield losses tend to increase with time. Thispaper reviews the research on the taxonomy, distribution, biology, pest status, sampling methods, and integratedpest management (IPM) of banana weevil. Salient features of the weevil’s biology include nocturnal activity, longlife span, limited mobility, low fecundity, and slow population growth. The adults are free living and most oftenassociated with banana mats and cut residues. They are attracted to their hosts by volatiles, especially followingdamage to the plant corm. Males produce an aggregation pheromone that is attractive to both sexes. Eggs are laidin the corm or lower pseudostem. The immature stages are all passed within the host plant, mostly in the corm. Theweevil’s biology creates sampling problems and makes its control difficult. Most commonly, weevils are monitoredby trapping adults, mark and recapture methods and damage assessment to harvested or dead plants. Weevil peststatus and control options reflect the type of banana being grown and the production system. Plantains and highlandbananas are more susceptible to the weevil than dessert or brewing bananas. Banana production systems rangefrom kitchen gardens and small, low-input stands to large-scale export plantations. IPM options for banana weevilsinclude habitat management (cultural controls), biological control, host plant resistance, botanicals, and (in somecases) chemical control. Cultural controls have been widely recommended but data demonstrating their efficacyare limited. The most important are clean planting material in new stands, crop sanitation (especially destructionof residues), agronomic methods to improve plant vigour and tolerance to weevil attack and, possibly, trapping.Tissue culture plantlets, where available, assure the farmer with weevil-free material. Suckers may be cleaned byparing, hot water treatment and/or the applications of entomopathogens, neem, or pesticides. None of these methodsassure elimination of weevils. Adult weevils may also invade from nearby plantations. As a result, the benefits ofclean planting material may be limited to a few crop cycles. Field surveys suggest that reduced weevil populationsmay be associated with high levels of crop sanitation, yet definitive studies on residue management and weevil peststatus are wanting. Trapping of adult weevils with pseudostem or corm traps can reduce weevil populations, butmaterial and labour requirements may be beyond the resources of many farmers. The use of enhanced trapping withpheromones and kairomones is currently under study. A combination of clean planting material, sanitation, andtrapping is likely to provide at least partial control of banana weevil.

Classical biological control of banana weevil, using natural enemies from Asia, has so far been unsuccessful.Most known arthropod natural enemies are opportunistic, generalist predators with limited efficacy. Myrmicine antshave been reported to help control the weevil in Cuba, but their effects elsewhere are unknown. Microbial control,using entomopathogenic fungi and nematodes tend to be more promising. Effective strains of microbial agents areknown but economic mass production and delivery systems need further development.

80 C.S. Gold et al.

Host plant resistance offers another promising avenue of control. Numerous resistant clones are known, includingYangambi-km 5, Calcutta 4, and Pisang awak. Resistance is most often through antibiosis resulting in egg or larvalfailure. Banana breeding is a slow and difficult process. Current research is exploring genetic improvement throughbiotechnology techniques including the introduction of foreign genes.

Neem has also shown potential for control of banana weevil. Studies on the use of other botanicals againstbanana weevil have failed to produce positive results. Chemical control of banana weevil remains a common andeffective method for larger scale producers but is beyond the reach of resource-poor farmers. However, the weevilhas displayed the ability to develop resistance against a broad range of chemicals.

In summary, cultural control remains the most available approach for resource-poor farmers. A combination ofseveral cultural methods is likely to reduce weevil pressure. Among the methods currently under study, microbialcontrol, host plant resistance and neem appear to offer the most promise.

Part 1: Biology and Pest Status ofBanana Weevil

The banana weevil, Cosmopolites sordidus (Germar),is an important pest of banana, plantain, and ensete.Weevil attack can prevent crop establishment, causesignificant yield reductions in ratoon cycles and con-tribute to shortened plantation life. For example, theweevil has been implicated as a primary factor con-tributing to the decline and disappearance of EastAfrican highland cooking banana (Musa spp., genomegroup AAA-EA) from its traditional growing areasin central Uganda (Gold et al. 1999b) and westernTanzania (Mbwana & Rukazambuga 1999).

The banana weevil is a difficult pest to work on.The adult is nocturnally active and seldom observed,while the immatures stages may be deep within thebanana corm. Damage often occurs well beneath thesoil surface. The insect’s biology creates a numberof sampling difficulties. Damage assessment requiresdestructive sampling that can affect the vigour and sta-bility of other plants on the mat. Adult population esti-mates are costly and there is only a modest relationshipbetween estimated adult densities and weevil damage.

All of the above factors have implications for theintegrated pest management (IPM) of banana weevil.The damaging larval stage is protected against mostnatural enemies by virtue of its cryptic lifestyle withinthe host plant. Control methods directed at the morevulnerable adult stage may not be directly translatedinto reductions in larval damage or may require con-siderable lag times before effects are felt. Controloptions requiring labour or costly inputs are depen-dent upon farmer objectives, management priorities,and allocation of limited resources.

Research results suggest that no single control strat-egy will be likely to provide complete control for

banana weevil. Therefore, a broad IPM approachmight provide the best chance for success in control-ling this pest. This paper provides a review of theavailable literature and unpublished data on bananaweevil biology, pest status, and management options.The paper concludes with recommendations for theway forward in developing management strategies forthis pest.

I. Banana Morphology and Phenology

The genus Musa evolved in southeast Asia (Stover &Simmonds, 1987). Edible bananas (Musa spp., Eumusaseries) originated from two wild progenitors, Musaacuminata and M. balbisiana, producing a series ofdiploids, triploids, and tetraploids through naturalhybridisation. Simmonds & Shepherd (1955) provideda key by which these naturally hybridised bananas maybe divided into six genome groups (AA, AAA, AAB,AB, ABB, ABBB) based on the relative contributionsof M. acuminata and M. balbisiana. Triploids tend tobe more vigorous and productive than diploids andcomprise the majority of currently cultivated bananas.Differences among these cultivars allow for differ-ent end products, i.e. dessert, cooking, roasting, andbrewing bananas. Some of these clones supply impor-tant international markets, while others are largelyrestricted to subsistence production or domestic trade.Bananas are grown from sea level to >2000 masl, undera range of different rainfall and soil conditions andin production systems ranging from kitchen garden tolarge-scale commercial plantations.

Bananas are herbaceous plants ranging in heightfrom 0.8 to 15 m (Turner 1994) that are vegetativelypropagated. A mat (=stool) consists of an undergroundcorm (rhizome) from which one or more plants (shoots)

Biology and IPM for banana weevil 81

emerge. The apparent stem or pseudostem is composedof leaf sheaths. The true stem arises from the api-cal meristem after leaf production has terminated andgrows through the centre of the pseudostem (Stover &Simmonds 1987). The true stem bears a single ter-minal inflorescence. After the fruit matures, the stemdies back to the corm. Farmers normally cut harvestedplants between ground level and 1 m.

New plants are produced by suckers emerging fromlateral buds in the corm. These can be left in situ(i.e. ratoon crops) or serve as a source of plantingmaterial, in which case they are removed and plantedelsewhere. Plant density is controlled by desucker-ing. Normally, a banana mat consists of three or moreplant generations (=ratoons or crop cycles) at any onetime. As banana stands age, mats ‘divide’ and the rela-tionship between plants (e.g. sharing of a commoncorm) becomes more tenuous; thus, in older stands,mat definition becomes unclear. Suckers used to estab-lish new fields are called the mother plant or plantcrop (Stover & Simmonds 1987; Turner 1994). Inolder stands, bananas are harvested throughout the year.Yield is normally expressed in kg/area/year and reflectsboth the number and size of bunches harvested.

II. Banana Weevil Taxonomy and Morphology

The banana weevil was first identified by Germarin 1824 from specimens collected in Java and giventhe name Calandra sordida. In 1885, Chevrolatchanged this to its currently recognised nameC. sordidus (Germar) (Viswanath 1976). Sphenophorusstriatus Fahreus 1845 (collected in Brazil) andS. cribricollis Walker 1859 (collected in Ceylon)are considered synonyms of C. sordidus and thesenames have been suppressed (Zimmerman 1968a,b).The genus Cosmopolites belongs to the subfam-ily Rhynchophorinae of the family Curculionidae(weevils and snout beetles). A single congenericspecies, C. pruinosus, is associated with bananas inIndonesia, the Philippines and the Caroline Islands(Zimmerman 1968b,c). Taxonomic keys are presentedby Zimmerman (1968a), while adult morphologyhas been described by Moznette (1920), Beccari(1967), Zimmerman (1968b), Viswanath (1976), andNahif et al. (1994), reproductive system morphologyby Cuille (1950), Beccari (1967), Uzakah (1995),and Nahif (1998, 2000), and larval morphologyby Moznette (1920) and Viswanath (1976). The

ultrastructure of the spermatazoan has been describedby Lino Neto & Dolder (1995).

The banana weevil’s limited mobility suggests theexistence of discrete populations with limited gene flowand the likely evolution of local biotypes. Studies onthe possible biotypes of banana weevils are currentlybeing concluded at ICIPE in Nairobi, Kenya (Ochieng2001; Ochieng et al. unpubl. data). Genetic diversityof banana weevils from East and West Africa, Asia,Australia, and the Americas were compared using ran-dom amplified polymorphic DNA polymerase chainreaction (RAPD-PCR) with five universal primers and46 RAPD markers. The data show that considerablevariation exists between banana weevil populationsfrom different parts of the world. Populations withgreatest levels of genetic similarity often came fromgeographically disparate areas (e.g. East Africa and theCaribbean), while populations from the same regionsometimes showed high levels of genetic diversity.Further examination of populations from different partsof Uganda showed distinct genetic variability, althoughmuch less than that found across regions.

III. Origin and Distribution

The banana weevil is believed to have originated in theIndo-Malayan region (Simmonds 1966; Zimmerman1968b; Waterhouse 1993), coincident with the areaof origin of bananas (Stover & Simmons 1987). Theweevil has since spread to all major banana-growingregions of the world, presumably through the move-ment of infested planting material. By 1900, theinsect was reported in Indonesia, China, Australia,and Brazil (Table 1). Within 20 years, it was alsoreported in sub-Saharan Africa, Central America,Pacific, and the Caribbean (Simmonds 1966). It iscurrently found throughout Asia, Oceana, Australia,Africa, and the Americas and absent only from banana-growing regions of North Africa (Cuille & Vilardebo1963).

In many countries, the banana weevil was prob-ably established long before it was first recorded.For example, bananas are believed to have enteredsub-Saharan Africa by multiple introductions betweenthe first and sixteenth centuries A.D. (Price 1995a;Karamura 1998). It is likely that the weevil entered theregion well before the first reports of its presence in theearly 1900s. Similarly, the weevil was first observed inCuba in 1944 although it was believed to have arrivedmany years earlier (Roche 1975).

82 C.S. Gold et al.

Table 1. First reports of the banana weevil C. sordidus in differentcountries

Region/Country Year Source

AsiaIndonesia 1824∗ Zimmerman (1968a)Sri Lanka 1859∗ Zimmerman (1968a)Sri Lanka 1885 Chevrolat (1885) in

Viswanath (1976)China 1885 Chevrolat (1885) in

Viswanath (1976)Vietnam 1885 Chevrolat (1885) in

Viswanath (1976)Taiwan 1909 Tsai (1986)India 1914 Moznette (1920)Malaysia 1914 Jepson (1914)Vietnam 1914 Jepson (1914)Philippines 1916 Moznette (1920)

OceanaFiji 1908 Moznette (1920)Borneo 1914 Jepson (1914)New Guinea 1914 Jepson (1914)Seychelles 1914 Jepson (1914)Guam 1936 Gressitt (1954)Hawaii 1981 Gettman et al. (1992)

AustraliaQueensland 1896 Franzmann (1976)New South Wales 1916 Hely et al. (1982)

AfricaMadagascar 1903 Moznette (1920)Sao Tome 1907 Gravier (1907)Uganda 1908 Hargreaves (1940)Congo (Zaire) 1913 Ghesquierre (1925)Tanzania 1922 Harris (1947)South Africa 1924 Cuille (1950)Sierra Leone 1925 Cuille (1950)Cote d’Ivoire 1938 Cuille (1950)French Guinee 1938 Cuille (1950)Canary Islands 1945 Carnero et al. (2002)Cameroon 1947 Nonveiller (1965)

AmericasBrazil 1845∗ Zimmerman (1968a)Brazil 1885 Arleu et al. (1984)Guadeloupe 1889 Harris (1947)Lesser Antilles 1912 Moznette (1920)Peru 1914 Jepson (1914)Dominican Republic 1916 Moznette (1920)Jamaica 1916 Moznette (1920)Guadeloupe 1916 Moznette (1920)Trinidad 1916 Moznette (1920)USA (Florida) 1917 Moznette (1920)Puerto Rica 1921 Wolcott (1948)Cuba 1944 Roche (1975)Colombia 1947 Gallego (1956),

Cardenas (1983)

∗From type specimens.

Nevertheless, surveys do suggest recent spread ofthe pest in some areas. For example, the weevilis believed to have been absent from the importantbanana-growing region of Bukoba district, Tanzania

until 1940 (Harris 1947) and from Kabarole district,Uganda until after 1957 (Whalley 1957; Gold et al.1993). Similarly, it was first recorded in the Colombiandepartments of Caldas in 1979, Quindio in 1981, andCesar in 1982 (Castrillon 1991, 2000). In the depart-ment of Risaralda, the banana weevil was found onnearly all farms during a survey in 1999 whereas it hadbeen found on only four farms in 1976 and on only 12%of farms between 1977 and 1981 (Castrillon 2000).Hargreaves (1940) estimated it would take 10 years forthe weevil to achieve pest status following its arrival in aregion.

Where present, severity of attack may be influ-enced by ecological conditions, clonal susceptibility,and management practices. Froggatt (1928) believedthat the hot weather conditions of low elevations inthe banana-growing zones of Australia reduced weevilactivity. However, the weevil thrives in the hot, humidconditions of coastal Nigeria (C. Gold pers. observ.)and Honduras (Sponagel et al. 1995).

Valentine & Valentine (1957) found the bananaweevil to be abundant at low elevations in Haiti butdid not observe it over 1200 masl. Lescot (1988) sur-veyed 45 sites and found a negative correlation (r =−0.75) of weevil damage with elevation with the great-est damage below 1000 masl and very low damagebetween 1500 and 1600 masl. The weevil was absentabove 1600 masl. Lescot (1988) also cited similar ele-vation thresholds in Burundi, Rwanda, and Colombia.In a later survey conducted in Colombia, the weevilwas most important at 1300–1400 masl (Anonymous1992).

However, in the Department of Risaralda, Colombia,Castrillon (2000) captured 1244 (2.5/trap) and 684(1.4/trap) banana weevils in plantain stands on farmsat 1600 and 1700 masl, respectively. This is the onlyrecord in the literature of significant weevil populationsabove 1600 masl. At three other sites above 1600 maslin this study, trap captures ranged from 0 to 0.2 weevilsper trap.

In Uganda surveys, weevil damage to highlandbanana clones (AAA-EA) was most severe between1000 and 1300 masl, although considerable variabil-ity existed at each elevation level (Gold et al. 1994a).Damage was not observed at the two sites above1600 masl. Gold & Okech (unpubl. data) have sincetrapped banana weevils (0.1–0.6 weevils/trap) andobserved low levels of damage on 10 farms between1600 and 2000 masl in Mbarara district, Uganda.

The upper elevation threshold for banana weevilis likely to be temperature related. Cuille (1950),Mesquita & Alves (1983), and Lescot (1988) suggest


minimal thermal thresholds for adult activity at15–18◦C, while thermal preferences have been esti-mated at 25◦C (Cuille 1950), 23–26◦C (Minost 1992),and 20–30◦C (Gomes 1985). In a controlled study,Traore et al. (1993, 1996) found minimal thermalthresholds of 12◦C for eggs and 10◦C for larvaewith highest rates of eclosion and larval develop-ment between 25◦C and 30◦C. These data suggest thatextended periods with low night-time temperatures athigher elevations may become bottlenecks for larvaldevelopment and/or adult survival. In Cameroon, forexample, nocturnal temperatures fall below <12◦C at1300 masl (Lescot 1988).

IV. Host Range

The banana weevil is a narrowly oligophagouspest, attacking wild and cultivated clones in therelated genera Musa (banana, plantain, abaca) andEnsete. Reports of alternative hosts, including sugarcane (Saccharum officinarum L.), yams (Dioscoreabatatas Dene), cocoyam (Xanthosoma sagittifolium(L.) Schott) (summarised by Moznette 1920; Beccari1967; Arleu & Neto 1984), have not been substanti-ated and appear to be in error. For example, gravidfemales placed in pots with Colocasia esculenta Schott,X. saggittifolium, and X. violaceum Schott failed tooviposit, while larvae inserted into these plants quicklydied (Martinez & Longoria 1990). Nevertheless, Traore(pers. comm.) was able to maintain larvae throughseveral instars using a factitious host (processedX. saggittifolium), while Pavis (1988) and Schmitt(1993) had modest success rearing larvae on artificialdiets.

V. Adult Biology

1. Longevity, tropisms, feeding, and distribution

The banana weevil displays a classical ‘K’ selectedlife cycle (Pianka 1970) with long life span and lowfecundity. Adults have been widely reported to liveupto 2 years (Froggatt 1925; Harris 1947; Nonveiller1965; Beccari 1967; Crooker 1979; Waterhouse &Norris 1987; Treverrow et al. 1992). In Uganda, markedweevils were recovered in experimental trials 4 yearsafter release (Rukazambuga & Gold unpubl. data).

Mean longevity under field conditions is not clear.Jardine (1924) reported that the adult lived 5–8 months,while Froggatt (1925) thought there was high mortality

soon after emergence. In Thailand, Jirasurat et al.(1989, cited in Vittayaruk et al. 1994) estimated maleand female longevity at 88 and 128 days, respectively.

The banana weevil adult is nocturnally activeand characterised by negative phototropism, stronghygrotrophism, thigmotactism, gregariousness, anddeath mimicry (Jardine 1924; Delattre 1980; Ittyeipe1986; Tsai 1986; Treverrow & Bedding 1993;Uzakah 1995; Braimah 1997; Aranzazu et al. 2000;Padmanaban et al. 2001). The adults have been widelyreported to favour crop residues and moist environ-ments, including in or under newly cut or rottingpseudostems, decaying stalks, cut or damaged corms,moist trash, and burrowed under the soil surface(Weddell 1945; Swaine 1952; Whalley 1957; Vilardebo1960, 1973; Nonveiller 1965; McNutt 1974; Roche &Abreu 1983; Jones 1986; Pavis 1988; Treverrow et al.1992; Silva & Fancelli 1998; Aranzazu et al. 2000).Additionally, the adults can penetrate the soil to depthsof 50–70 cm (Cardenas & Arango 1986; Rukazambugaunpubl. data). Moznette (1920), Vilardebo (1960),Saraiva (1964), Treverrow et al. (1992), and Silva &Fancelli (1998) reported adults to be closely associ-ated with the banana mat, being primarily in the leafsheaths, around the roots, under loose fibres surround-ing the base of the plant and, occasionally, in larvalgalleries. Silva & Fancelli (1998) reported that, duringthe day, the adults may also be found in wet, shadedareas under shrubs.

In Uganda, banana plots were systematicallysearched to determine field distribution of bananaweevil adults (Gold et al. 1999d). Most banana weeviladults were within or attached to the banana plants(mainly in leaf sheaths) (41%), or in the soil at thebase of the mat (24%) (Gold et al. 1999d). Weevil den-sity was greatest in or around flowered plants, althoughadults were also commonly found on pre-floweredplants and harvested stumps. Significant numbers ofweevils (28%) were attached to cut and prostrateresidues >0.5 m away from the mats, while negli-gible numbers were found in the leaf litter (6%) orburied in soil >0.5 m from the banana mat (1%). A fewmarked weevils had re-entered the corm or pseudostemof living plants through existing galleries. Distributionpatterns of males and females were similar.

The adults feed on rotting banana tissue (Budenberget al. 1993b) or, occasionally, on young suckers(Castrillon 2000), but it is likely that resulting dam-age is negligible. The weevil can survive without foodfor extended periods (2–6 months) in moist environ-ments (Moznette 1920; Froggatt 1924, 1925; Cuille &Vilardebo 1963; Viswanath 1976; Crooker 1979;

84 C.S. Gold et al.

Mitchell 1980; Aranzazu et al. 2000; Castrillon 2000).In one exceptional case, females were maintained for17 months without food (Franzmann 1976). Underfield conditions, the weevil reportedly can survive3–6 months once all banana plants and residues areremoved (Froggatt 1924; Peasley & Treverrow 1986;Allen 1989).

However, the weevil is very susceptible to desic-cation and will die within 1–10 days if kept in adry substrate (Froggatt 1925; Cuille 1950; Vilardebo1960; Cuille & Vilardebo 1963; Viswanath 1976; Gold1998a). For example, Viswanath (1976) maintainedweevils in moist soil without food for 112 days, butthe weevils died in 10 days when kept in dry soil. As aresult, the weevil is strongly hygrotropic and searchesfor the highest available air humidity and for liquidwater (Cuille 1950; Roth & Willis 1963). In labora-tory studies, males and females were unsettled at lowhumidity and sedentary at high humidity (Roth & Willis1963). Koppenhofer (1993a) suggested that femalesdeposit eggs on living plants further below the soilsurface during the dry season. Masanza (unpubl. data)found a higher proportion of oviposition on buriedcorms (i.e. chopped below the soil surface) in the dryseason than in the wet season.

2. Sexual dimorphism and sex ratio

Males can be distinguished from females on the basis ofpunctuation on the rostrum extending beyond the pointof insertion of the antennae (McCarthy 1920 cited inLongoria 1968; Viswanath 1976), a shorter and lessaccentuated curvature of the rostrum (Longoria 1968;Mestre 1995) and greater curvature of the last abdom-inal sternite (Roth & Willis 1963). Males tend to beuniform black or dark-brown, while the rostrum offemales may be redder in colour than the rest of theirbody (Longoria 1968). Punctuation of the rostrum andcurvature of the last abdominal sternite appear to bethe most reliable methods for sexing weevils and arewidely used. In Uganda, dissections suggested >95%accuracy in sexing weevils on the basis of these traits(Rukazambuga et al. 2002; Abera & Gold unpubl.data).

The range in adult size has been reported as10–16 mm (Nonveiller 1965), 8.8–13.2 mm (Beccari1967), 11–14.5 mm (Viswanath 1976), 11–14 mm(Sponagel et al. 1995; Carnero et al. 2002), and15–20 mm (Aranzazu et al. 2000, 2001; Castrillon2000), although mean sizes of 12.5–13.0 mm weresimilar in such disparate populations as those in

India (Viswanath 1976), Cameroon (Lescot 1988),and Honduras (Anonymous 1989; Sponagel et al.1995). On average, females are >20% longer (Cuille1950; Beccari 1967; Sponagel et al. 1995) andweigh 11–17% more than males (Edge 1974; Goldet al. 1999a,d).

Sex ratios (female : male) of field-collected weevilshave been reported as 1 : 1 in Guinee (Cuille 1950),Kenya (Koppenhofer & Seshu Reddy 1994), and theCanary Islands (Carnero et al. 2002), 1.2 : 1 in Tonga(Litsinger 1974), 1 : 1.4 in India (Viswanath 1976),and 1 : 2.2 in Honduras (Sponagel et al. 1995). InCameroon, Delattre (1980) found a sex ratio of 1 : 1 forlaboratory-reared weevil and attributed a higher pro-portion of females among field-trapped weevils in therainy season to sexual differences in behaviour. In asurvey of 50 farms in Ntungamo district, Uganda, theoverall sex ratio for trapped weevils (N = 15,376) was1.07 : 1, although on individual farms it ranged from1.67 : 1 to 1 : 1.56 (Gold et al. 1999d; Gold & Okechunpubl. data).

3. Daily and seasonal activity periods

Banana weevil adults are negatively phototropic andtend to be sedentary during daylight hours. They areactive between 1800 and 0600 hours (Cuille 1950;Uzakah 1995) with greatest activity between 2100 and0400 hours (Uzakah 1995). Padmanaban et al. (2001)collected weevils in freshly set traps during daytimehours in mulched fields suggesting at least some diur-nal weevil activity in these systems. A substantial pro-portion of the population may be inactive for extendedperiods (S. Lux pers. comm.).

Seasonal differences in trap captures have beenreported by many authors. Trap captures, however, donot provide meaningful estimates of population den-sity (Vilardebo 1973), which require mark and recap-ture methods (Price 1993; Gold & Bagabe 1997). Mostlikely such apparent seasonal differences reflect weevilactivity patterns more than population fluctuations.

In Latin America (primarily Brazil and Cuba), trapcaptures reported in a number of studies suggestreduced adult activity in the rainy season (Yaringano &van der Meer 1975; Reinecke 1976; Zem & Alves1976; Mesquita et al. 1981; Arleu & Neto 1984 (onestudy); Gomes 1985; Durans Pinheiro & Batista deCarvalho Filho 1985; Bendicho & Gonzales 1986;Batista Filho et al. 1992). In contrast, other researchers(including most African and some Latin Americanstudies) found greater weevil activity and trap captures


in the rainy season (Cuille 1950; Roy & Sharma 1952;Whalley 1957; Cuille & Vilardebo 1963; Saraiva 1964;Delattre 1980; de Souza et al. 1981; Vilardebo1984; Marcelino & Quintero 1991; Pinese & Piper1994; Price 1995b). Still others found no relationshipbetween climatic factors (rainfall, relative humidity,and/or temperature) and trap captures (Sen & Prasad1953; Oliveira et al. 1976; Delattre 1980; Arleu 1982;Pulido 1982, 1983; Arleu & Neto 1984 (second study);Van den Enden & Garcia 1984; Cardenas & Arango1986; Pavis 1988; Boscan de Martinez & Godoy1989). Uzakah (1995) reported activity in the labo-ratory to be positively correlated to relative humidityand negatively correlated with temperature and lightintensity. These conflicting results provide an unclearpicture of when adults are most active and, hence, mostvulnerable to control interventions.

Banana weevil populations may show year toyear fluctuations reflecting environmental conditions(e.g. drought). For example, in a study evaluating thecontrol potential of pseudostem trapping, weevil pop-ulations on nine control farms declined from a mean of13,400/ha to 11,400/ha (farm mean change = −38%)in 1 year (Gold et al. 2002b). In this study, however,populations declined by 42% during the first 6 monthsin which rainfall was 635 mm and then increased by12% during the next 6 months during which rainfallwas 245 mm (Gold et al. unpubl. data).

4. Dispersal and movement

a. CrawlingDispersal by means of crawling appears to be lim-ited and slow. Moznette (1920) reported most adultsto remain near their sites of emergence. Delattre(1980) found 90% of weevils recaptured after 3 daysto be at the point of release. Whalley (1957) andCardenas & Arango (1986) each reported that mostweevils moved less than 10 m over a period of sev-eral months. Maximum weevil movement has beenrecorded as 6 m in a night (Wallace 1938), 15 m in anight (Cendana 1922), 21 m in 14 days (Wallace 1938),35 m in 3 days (Gold & Bagabe 1997), and 60 m in5 months (Delattre 1980). Most of these trials con-cerned tracking weevils released at a single point, withprobability of recapture declining at greater distancesfrom the release point.

Wallace (1938) found few weevils were able tocross grass barriers of 4–10 m. In a series of trialslasting several years and in which plots (15 × 15 m2

and 15 × 25 m2) were separated by 20 m grass alleys,

Gold et al. (1998b, unpubl. data) found that less than3% of marked weevils appearing in pseudostem trapswere captured in plots other than those in whichthey had been released. Nevertheless, Vilardebo (1984)found weevils attracted to house lights 200 m from thenearest banana stand.

Mestre & Rhino (1997) released five marked femalesand five marked males each in a series of pseudostemtraps, which they then monitored for 17 days. Thepercentage of marked weevils in these traps declinedrapidly; after 17 days, 25% of the released weevilshad remained in the traps. Mestre & Rhino (1997)observed much faster disappearance of females thatthey hypothesised was associated with their search foroviposition sites.

In a study at the Kawanda Agricultural ResearchInstitute near Kampala, Uganda, 2000 weevils wereindividually marked with distinctive patterns, releasedin banana stands and tracked by pseudostem trappingfor 1 year. Results from the first 10 weeks have beenpublished (Gold et al. 1999d). Two weeks after release,49% of the weevils in a mulched plot and 79% in anunmulched plot were found at the site of release. By10 weeks after release, 17% and 36% of the weevilswere recovered at the release point in mulched andunmulched plots, respectively. During the same timeperiod, 60% of the weevils had moved >10 m in themulch, while 27% had moved >10 m in the unmulchedplot. Six months after release, 42% of recapturedweevils were found within 5 m of the point of release,39% had moved 6–15 m and only 3% had moved morethan 25 m (Gold & Kagezi unpubl. data). These resultssuggest that adults may be sedentary for extendedperiods and that soil moisture stimulates activityand movement. Females tended to be more activethan males, leaving the site of release quicker andmoving longer distances (Gold & Kagezi unpubl.data).

b. FlightAlthough the banana weevil has functional wings, mostobservers report that the weevil seldom if ever flies(Froggatt 1925; Nonveiller 1965; Gordon & Ordish1966; Wardlaw 1972; Cardenas & Arango 1986;Greathead 1986; Waterhouse & Norris 1987; Pinese &Piper 1994; Sponagel et al. 1995). At ICIPE, flightwas never observed during laboratory studies on noc-turnal weevil behaviour (Uzakah 1995; S. Lux pers.comm.). In the laboratory, a few weevils were observedto open their wings in response to extreme drought, thelatest stages of insecticide influence (Whalley 1957)

86 C.S. Gold et al.

or being tethered (Viswanath 1976), but did not fly.Gold & Nankinga (unpubl. data) placed weevils on hotplates but could not induce flight by gradually increas-ing the temperature until the weevils died. Reports ofbanana weevil flight (Cuille & Vilardebo 1963; Haarer1964; Simmonds 1966) are often anecdotal.

Nevertheless, the possibility of dissemination byflight remains unclear. Few workers have made obser-vations on the weevil during it periods of greatestactivity (i.e. 2100–0400 hours) and flight may be stim-ulated only under certain environmental conditions.Castrillon (pers. comm.) believes that weevils read-ily fly at night following the removal of host plantsin recently uprooted banana stands. In Cameroon,Messiaen (2002) captured 14 adult banana weevilsover a 40-week period in window traps placed 0.5–1.5 m above the ground on the edge of a 0.8 habanana stand. Moreover, weevil attack of isolatedfields planted with tissue culture plants in Malaysiaand Uganda (C. Gold pers. observ.) suggests that dis-persal by flight may be greater than most observersbelieve.

c. DispersalIn summary, the data suggest that banana weevils arerelatively sedentary and dissemination by crawling orflight of adults is limited. Adult weevils can movewithin banana stands and across contiguous plantings.However, movement across barriers of >20 m may belimited. Moreover, the banana weevil’s narrow hostrange and limited dispersal capability mitigate againstimmigration of adults into isolated or newly plantedbanana stands (Gold et al. 1998b, 1999d).

It has been widely recognised that dispersal ofbanana weevil is primarily through infested plant-ing material (Froggatt 1925; Ghesquierre 1925;Hargreaves 1940; Vilardebo 1960; Haarer 1964;Nonveiller 1965; Cardenas & Arango 1986; Jones1986; Rodriguez 1989; Pinese & Piper 1994; Gold et al.1998a,b; Castrillon 2000). Infested planting material islikely to contain adults in the leaf sheaths and imma-ture stages in the pseudostem and corm. For exam-ple, Abera et al. (1999) reported 0.4–1.5 eggs per plant(cv Atwalira, AAA-EA) for peepers and suckers, whileGold et al. (1998a) found mean larval infestation levelsper sucker of 0.3 for Gonja (AAB), 0.6 for Atwalira,and 1.1 for Nsowe (AAA-EA). This suggests that theuse of clean planting material is an important factor inestablishing healthy banana stands and retarding weevilbuild-up.

5. Teneral stage

The teneral stage of newly emerged adults is mostoften passed in the pupal chamber within the plant(Jardine 1924; Longoria 1972). Teneral adults arereddish-brown and turn dark-brown to black as theirexoskeletons harden. Under tropical conditions, theteneral stage can last from 2 to 14 days (Jardine 1924;Froggatt 1925; Montellano 1954; Viswanath 1976;Mestre 1997). In contrast, Cuille (1950) described ateneral stage in which the weevil required 22–60 daysto reach its final black colour.

6. Mating

Mating is most often at night (Delattre 1980; Jones1986; Uzakah 1995). Courtship and mating behaviourof the banana weevil have been described by Uzakah(1995) and Viana & Vilela (1996). Mating lasted3–24 min (mean 7.5 min). Following mating, malesoften guard the females to prevent further mating. Theweevils may oviposit for up to 11 months withoutrenewed mating (Cuille 1950). Treverrow et al. (1992)reported production of up to 100 eggs following a singlemating.

7. Sexual maturity and preoviposition period

In Kenya, male sexual maturity (i.e. the ability toinseminate females) was attained at 18–31 days afteremergence (DAE) (Uzakah 1995). Spermatogenesisoccurred at emergence, but sperm were non-motilein the female following insemination. This suggestedthat secretions from an accessory gland activate sperm.Female sexual maturity was at 5–20 DAE, the firstoocytes were observed at 11–28 DAE, chorionatedeggs first appeared at 25 DAE, and first ovipositionoccurred at 27–41 DAE (Uzakah 1995). This suggeststhat oocytes require about 2 weeks to mature. Onlymated females produced chorionated eggs (Uzakah1995). In Tonga, 4% of the females were infertile(Litsinger 1974).

First oviposition has also been reported at 7–10 days(Treverrow & Bedding 1993), 21 days (Viswanath1976), 33–36 days (Cuille 1950), and >60 days(Pulido 1982). Froggatt (1925), Arleu (1982) andSilva & Fancelli (1998) reported greater oviposi-tion rates in young females, while Cuille (1950) andTreverrow et al. (1992) found that females more than1-year-old produced as many eggs as young females.


Viswanath (1976) observed maximum oviposition dur-ing the 7th month with greatly reduced oviposition after11 months.

8. Oviposition potential

The banana weevil has telotrophic ovaries with eachcontaining two ovarioles (Uzakah 1995; Nahif 1998),although a few individuals may have only one ovar-iole per ovary (Abera & Gold unpubl. data). Thereis normally continual development of oocytes inthe ovaries (Froggatt 1925), although in the labora-tory studies undernourished weevils ceased oviposit-ing and their ovaries may became non-functional(Cuille 1950; Cuille & Vilardebo 1963). Cuille (1950)found 4 oocytes per ovary and suggested that fooddeficiency arrests oviposition and leads to reabsorp-tion. According to Nahif (1998), each calyx canhold 4 eggs. In contrast, Uzakah (1995) and Abera(1997) reported up to 17 (mean 5) and 22 (mean 10)chorionated eggs, respectively, retained in thecalyces.

Dissections of 1140 females after being maintainedin the laboratory for 30 days on oviposition substrates,revealed an average of 1.7 mature eggs (range 0–16),2.0 medium-sized oocytes (range 0–10), and 4.9 smalloocytes (range 0–13) per weevil (Gold et al. 2002a,unpubl. data). The total number of eggs and oocytesaveraged 8.5 (range 0–24).

9. Realised oviposition

Egg production of the banana weevil is low, with ovipo-sition in the laboratory most commonly estimated at1–4 eggs/week (Cuille 1950; Vilardebo 1960, 1984;Cuille & Vilardebo 1963; Haarer 1964; Gordon &Ordish 1966; Delattre 1980; Pulido 1983; Arleu &Neto 1984; Treverrow et al. 1992; Minost 1992;Treverrow & Bedding 1993; Koppenhofer 1993a;Lemaire 1996) and 10–270 in the lifetime of the insect(Cuille 1950; Viswanath 1976; Arleu & Neto 1984;Castrillon 1989; Treverrow et al. 1992; Treverrow &Bedding 1993; Aranzazu et al. 2000, 2001). In con-trast, Montellano (1954) reported oviposition rates of 1and occasionally 2 eggs/day. Further laboratory studiesin Uganda found oviposition rates of 4–14 eggs/week(Rukazambuga 1996; Abera 1997; Griesbach 1999;Gold et al. 2002a). Females may also pass extendedperiods without any oviposition (Cuille 1950; Longoria1972; Cardenas 1983).

In India, Viswanath (1976) found a mean ovipo-sition of 43 eggs (range 36–53) in the lifetime ofthe weevil. The oviposition period averaged 471 days(range 313–556 days), suggesting 1 egg every 11 days.Maximum oviposition (4 eggs/month) occurred dur-ing the 7th month. Females more than 11 months oldproduced less than 1 egg/month. All oviposition wasat night. The post-oviposition period lasted 35 days(range 14–57 days).

Under field conditions near Kampala, Uganda,oviposition was estimated at 0.5–1.4 eggs/week (Abera1997). Seasonal effects on oviposition are unclear.Uzakah (1995) found oviposition rates related totemperature but not to relative humidity or rain-fall. However, Cuille (1950) and Cuille & Vilardebo(1963) reported oviposition of 7.8 eggs/female/monthin the rainy season and 0.4 eggs/female/month inthe dry season. Similarly, Nonveiller (1965) andSimmonds (1966) report reduced oviposition in the dryseason.

Although Uzakah (1995) found no relationshipbetween female size and egg production, Griesbach(1999) reported smaller weevils produce fewereggs. Griesbach divided field-collected banana weevilfemales into ‘large’ (mean weight 0.11 g) and ‘small’(mean weight 0.06 g) individuals. The large femaleslaid significantly more eggs (mean = 0.43/day) thansmall females (0.28/day) (T = 4.76; p < 0.01).Large weevils also produced significantly larger eggs(0.47 mg) with higher rates of eclosion (81%) than eggsproduced by small weevils (0.41 mg; 73%).

In a separate experiment, Abera et al. (unpubl. data)found that large and small field-collected weevils con-tained similar numbers of chorionated eggs (4.0 and4.3). When held in the laboratory for 2 or 6 weekswithout exposure to an oviposition substrate, largerweevils maintained twice as many chorionated eggs(10.5 and 11.3, respectively) as did small weevils (5.0and 4.6). These data suggest that weevils reabsorb eggsand oocytes and that the rate of reabsorption may begreater for smaller individuals.

In Uganda, available data suggest that weevils aremore active under conditions of higher soil moisture(i.e. rainy season or under mulches) and it is likely thatoviposition is greater at this time. Ovipisition rates mayalso be influenced by weevil density and temperature(Rukazambuga 1996; Silva & Fancelli 1998; Gold et al.2002a, unpubl. data). In Brazil, for example, Silva &Fancelli (1998) report higher oviposition rates at 24◦Cthan at 28◦C.

88 C.S. Gold et al.

In summary, weevil dissections indicate that femalesproduce four or more oocytes per week, have the capac-ity to store eggs until a favourable substrate is foundand can reabsorb eggs under unfavourable conditions.However, realised oviposition in the field may be con-siderably less than the weevil’s potential fecundity.Why this may be is unclear since banana stands con-tain an abundance of host substrate for oviposition. Lowoviposition rates may, in part, explain the slow build-upof weevil populations over time. It also suggests thatstrategies targeting adults may have a long-term impacton weevil numbers and subsequent damage.

10. Oviposition preferences: Timing of attack

Timing of oviposition with respect to host phenolog-ical stage has implications for understanding yieldloss (e.g. possible critical periods of attack), screen-ing methods for host plant resistance (larval successmay vary by age of plant), timing of interventions andstrategies for managing crop residues (i.e. sanitation).

Banana weevils oviposit on all stages of bananaplants ranging from peepers (i.e. newly emergedsuckers) to crop residues (Abera et al. 1999). Femalesare attracted to freshly cut corms making youngsuckers, recently detached from mother plants, espe-cially vulnerable. Suckers planted in or proximal toinfested fields may fail to establish due to weevil attack(McIntyre et al. 2002).

Vilardebo (1973, 1984), Haddad et al. (1979) andMesquita & Caldas (1986) suggested that oviposit-ing weevils favour plants during flowering andbunch maturation. In the laboratory, Cerda et al.(1995) found weevils oriented more towards cormsof flowered than of pre-flowered plants. In contrast,Cuille (1950) reported that the weevil prefers tooviposit on young plants, while Treverrow & Maddox(1993) noted heavy attack on pre-flowered plants.Ittyeipe (1986) suggested that ovipositing weevilsdo not discriminate among bananas on the basis ofplant age. In a field trial in Uganda (cv Atwalira,AAA-EA), weevils released at a density of 20 per matoviposited on 23% of peepers (0.6 eggs/plant), 35% ofmaiden suckers (1.2 eggs/plant), 74% of pre-floweredplants (6.0 eggs/plant), and 90% of flowered plants(13.8 eggs/plant) (Abera et al. 1999). Egg density perunit surface area was 2.5–4 greater on flowered plantsthan earlier stages.

The banana weevil will also oviposit in crop residues(Froggatt 1925; Vilardebo 1960; Koppenhofer 1993a;Gold & Bagabe 1997; Abera et al. 1999). Prostrate

stems are considered favoured breeding grounds forthe weevil (Treverrow & Maddox 1993). Abera et al.(1999) found oviposition on 88% of the residues of har-vested highland banana plants (19 eggs/plant), whileRukazambuga (unpubl. data) collected 200 eggs froma single stump (cv Atwalira).

In Uganda, the cultivar Kisubi (AB, Ney Poovan sub-group) is resistant to banana weevil attack, yet high lev-els of damage may be found in crop residues (more soon prostate rather than standing stems) (Gold & Bagabe1997). In Indonesia, Gold & Hasyim (pers. observ.)found up to 100 larvae on prostrate stems, while stand-ing and recently harvested stumps were virtually freeof attack. Abera (1997) and Kiggundu (2000) sug-gested that damage on residues of resistant clonesreflect larval success rather than ovipositional prefer-ences. This would imply the breakdown of biochemical(antibiotic) defences following harvest. Greater suc-cess on prostrate stems may also reflect exposure of thetrue stem to ovipositing females. Banana weevil larvaeprefer to feed on the corm and true stem and are uncom-monly found feeding in the pseudostem. In contrast,young weevil larvae on stumps and unharvested plantsoften have to tunnel from oviposition sites through thepseudostem to reach their preferred feeding sites.

Montellano (1954) found eight times as many eggson fresh versus decomposing corms, while tissues instates of advanced deterioration were rejected entirely.However, Masanza (unpubl. data) found oviposition onmoist residues up to 120 days after harvest, althoughmost oviposition was on residues less than 30 days old.

11. Oviposition sites

Eggs (0.5 × 2 mm2) are deposited singly in the hostplant in orifices (1–2 mm deep) excavated by thefemale weevil with her rostrum. Oviposition sites havebeen described by many authors although few stud-ies have quantified egg distribution within the plantunder field conditions. In general, it is believed thatoviposition is usually at the base of the plant, at ornear soil level, in the leaf sheaths at the base of thepseudostem (Jepson 1914; Moznette 1920; Jardine1924; Pinto 1928; Harris 1947; Swaine 1952;Montellano 1954; Viswanath 1976; Arleu 1982;Suplicy Fo & Sampaio 1982; Arleu & Neto1984; Pinese & Piper 1994; Pavis & Lemaire 1997;Abera 1997), in leaf scars (Saraiva 1964; PANS1973; Treverrow 1985; Allen 1989; Waterhouse &Sands 2001), the corm (Pinto 1928; Montellano1954; Beccari 1967; Koppenhofer 1993a; Abera 1997;


Silva & Fancelli 1998; Nkakwa 1999; Castrillon 2000;Gold & Kagezi unpubl. data), and in the weevil gal-leries in the interior of the corm (Montellano 1954;Martinez & Longoria 1990; Koppenhofer 1993a).Oviposition on roots is uncommon (Abera 1997).

There is some disagreement whether eggs are placedabove or below the soil surface. Oviposition place-ment may be influenced by weather, plant stage, thepresence or absence of high mat and the availabilityof crop residues (Koppenhofer 1993a; Abera 1997).The location of eggs will affect vulnerability to nat-ural enemies (Koppenhofer 1993a); those below thesoil surface are likely to be relatively protected againstpossible parasitoids and predators. In laboratory exper-iments, Cuille & Vilardebo (1963) found 69% of theeggs on the corm with the remainder on the pseu-dostem. Of those in the corm, 16% were in basal third,31% in the middle, and 53% in the upper third. In pottrials, Koppenhofer (1993a) found egg distribution tobe 33% in crown area, 5% superficially in base of roots,29% inside of abandoned tunnels, 22% in remainingparts of corm, and 11% in base of leaf sheaths. Themajority of eggs were below the soil surface, with bothadults and eggs found to a depth of 50 cm.

In field studies, Abera (1997) found 96% of eggson plants without high mat to be in the leaf sheaths,4% in the corm and 1% on the roots. Seventy-fivepercent of the eggs were placed below the soil surface.Oviposition reached 15 cm above the collar although60% of the eggs on the pseudostem were <5 cm abovethe collar. High mat increased both total oviposition,the proportion of eggs on the corm and the percent-age of eggs placed above the ground (Abera 1997).Koppenhofer (1993a) estimated that 50% of the eggswere accessible to predators, while Abera’s (1997) datasuggest that a lower percentage of eggs may actuallybe vulnerable to predation.

Crop residues are also attractive to ovipositingweevils. On prostrate stems, most eggs are placedwithin 12–18 inches of basal end or, if part of the cormis attached, just above the crown (Froggatt 1925). Mosteggs placed on residues are probably vulnerable tonatural enemy attack.

VI. Development of Immatures

Banana weevil developmental rates determined underambient temperatures (reviewed by Schmitt 1993;Traore et al. 1993) show wide variability in stageduration: 3–36 days for eggs, 12–165 days for larvae,

1–4 days for prepupae, 4–30 days for pupae, and24–220 days from egg to adult. The longest stage dura-tions were found in Australia where seasons are pro-nounced and the range in weevil development timeslarge: In cold seasons, development rates were up tofour times as long as recorded anywhere else. Whiletemperature is certainly the most critical factor in deter-mining developmental rates, relative humidity, cultivar,age of plant, food quality, and population density mayalso be involved (Mesquita et al. 1984; Schmitt 1993).

1. Egg stage duration and eclosion rates

Most studies on egg stage duration have been con-ducted under ambient temperatures. Under tropicalconditions, the egg stage has been most commonlyfound to last 6–8 days (Pinto 1928; Cuille1950; Montellano 1954; Vilardebo 1960; Saraiva1964; Woodruff 1969; Longoria 1972; Viswanath 1976;Mesquita & Alves 1983; Pinese & Piper 1994;Vittayaruk et al. 1994; Seshu Reddy et al. 1998; Goldet al. 1999d), although some researchers reported anegg stage of 8–10 days (Montellano 1954; Vilardebo1960; Ingles & Rodriguez 1989; Padmanaban et al.2001). However, eclosion may occur in as few as3–4 days (Froggatt 1924; Cuille 1950; Longoria 1972;Franzmann 1976; Mesquita & Alves 1983; Pinese &Piper 1994) or as many as 14–15 days (Montellano1954; Vilardebo 1960; Trejo 1969; Mesquita & Alves1983).

In Cotonou, Benin (mean temperature 26.8◦C),Traore et al. (1993) monitored egg stage durationunder six constant temperatures ranging from 15◦Cto 34◦C and determined a developmental thresholdof 12◦C and thermal requirement of 89 degree-days.The duration of the egg stage decreased from 35 daysat 15◦C to 5 days at 30◦C. Highest rates of eclo-sion occurred between 25◦C and 30◦C. Eggs did nothatch above 32◦C. In Brazil, Ferreira (1995) deter-mined egg stage duration to be seven to nine days (mean8.0 days) at 25◦C.

Local biotypes may exist and immatures may displayoptimal development at the prevailing temperatures fora given site. Nevertheless, estimates of degree-day ther-mal requirements for banana weevil eggs in Kampala,Uganda (mean temperature 22.5◦C) (Gold et al. 1999c)suggested conformity to developmental periods estab-lished for West African banana weevil populations byTraore et al. (1993).

Eclosion rates of over 80% have been recorded(Bakyalire 1992; Griesbach 1999) and it is likely that

90 C.S. Gold et al.

field eclosion on susceptible cultivars may approach100%. However, in the laboratory, egg mortalitymay be very high (Gomes 1985; Minost 1992;Ogenga-Latigo 1992; Traore et al. 1993; Treverrow &Bedding 1993; Lemaire 1996; Carnero et al. 2002)due to handling, desiccation, and fungal attack.Koppenhofer & Seshu Reddy (1994) found lowerhatchability for eggs in pseudostems, possibly dueto higher water content or metabolites. Kiggundu(2000) suggested that viscosity and metabolites ofplant sap in resistant clones might also reduce eggsuccess.

2. Larval stage

a. Distribution, feeding, and damageFirst-instar larvae emerging in the leaf sheaths tendto move downward into the corm. Jardine (1924),Vilardebo (1960), Ittyeipe (1986), and Sponagel et al.(1995) suggested that the larvae prefer the cortical tis-sue to the central cylinder. In Ugandan surveys, Goldet al. (1994b) estimated damage to the central cylin-der and cortex in 7000 recently harvested plants andfound that 43% of weevil damage in plantain was in thecentral cylinder compared to 36% in highland banana,26% in Pisang awak, 22% in Gros Michel, and 17% inAB clones. Within the highland banana group, the pro-portion of damage found in the central cylinder rangedfrom >40% of total damage in Namwezi, Muskala,and Nakitembe to 25% in Kibuzi, Mbwazirume, andNakyetengu.

The larvae feed throughout the corm but there islimited information on their vertical distribution rel-ative to the soil surface and distance below the col-lar. In the Ugandan surveys, Gold et al. (unpubl. data)consistently found greater damage at 10 cm belowthe collar than at the collar. Englberger & Toupu(1983), Price (1994), and Gold & Kagezi (unpubl. data)found highest levels of damage on the lower third ofthe corm.

The larvae may also enter the true stem after flower-ing. In severe attacks, the larvae may feed on leaf tis-sue in the pseudostems or move from the mother plantinto young suckers (Vilardebo 1960; Champion 1975;C. Gold pers. observ.). In exceptional circumstances,larval galleries can reach 30–100 cm above the collar(Moznette 1920; Sen & Prasad 1953; Treverrow 1985).In one case, Froggatt (1925) observed three larvae inthe stalk of the fruit.

Measuring gallery size is difficult because ofthe larva’s meandering feeding habit within the

corm. Beccari (1967) suggested that first-instar lar-vae tunnel 7–8 cm before moulting. Maximum gallerydiameter has been reported to range from 0.8 cm(Sponagel et al. 1995; Seshu Reddy et al. 1998) to1.2 cm (Montellano 1954), while gallery length hasbeen variously reported as 30 cm (Treverrow 1985),60 cm (Cuille 1950), 63 cm (Montellano 1954), 70 cm(Beccari 1967), 120 cm (Sponagel et al. 1995), and150 cm (Seshu Reddy et al. 1998). The wide range inreported gallery size makes it difficult to estimate thenumber of larvae that cause observed levels of damagein a corm.

b. InstarsThe banana weevil has been variously reported tohave 5 (Cendana 1922; Beccari 1967; Lemaire 1996;Carnero et al. 2002), 6 (Cuille 1950; Montellano 1954;Koppenhofer et al. 1994), 7 (Viswanath 1976; Pulido1982; Ferreira 1995), 4–6 (Traore et al. 1996), 4–7(Mestre 1997), 5–7 (Schmitt 1993), 5–8 (Mesquitaet al. 1984; Mesquita & Caldas 1986; Gold et al.1999c), or 6–7 instars (Cuille 1950; Arleu & Neto1984). The variable number of larval instars in somestudies suggest that banana weevils may display devel-opmental polymorphism, i.e. the occurrence of instarnumber other than those which are thought to be‘customary’ for a particular species (c.f. Schmidt &Lauer 1977).

Developmental polymorphism in banana weevil hasbeen attributed to temperature (Schmitt 1993; Traoreet al. 1996), nutrition (Cuille & Vilardebo 1963), clone(Haddad et al. 1979; Mesquita et al. 1984; Mesquita &Caldas 1986), plant stage (Mesquita et al. 1984;Mesquita & Caldas 1986), and rearing method (Goldet al. 1999c). Adverse conditions or resistant clonesincreased the numbers of moults, prolonged stage dura-tion and resulted in smaller pupae (Mesquita & Caldas1986; Gold et al. 1999c).

Gold et al. (1999c) set up frequency distributionpolygons of head capsule-widths to separate instarsfor laboratory-reared and field-collected larvae. Meanhead capsule-widths for the first four instars showedclose agreement among laboratory-reared and field-collected populations. The method of analysis was notsensitive enough to separate later instars. However,field-collected larvae were larger than those reared inthe laboratory, thus forming distinct frequency distri-bution curves. Traore (1995) and Sponagel et al. (1995)also present ranges of head capsule sizes for differentinstars but did not attempt to fit these sizes to frequencydistribution curves.


c. Larval stage durationUsing weevils collected in southern Benin and Onne,Nigeria, Traore et al. (1996) determined developmentalthresholds and thermal requirements for each instar forlarvae kept under five constant temperatures rangingfrom 16◦C to 30◦C. The total larval period (includingprepupal stage) was inversely related to temperatureranging from 34 days at 30◦C to 70 days at 16◦C.Development in each of the six observed instars showeda similar inverse relationship with temperature. Thedevelopmental threshold for the larval stage was 8.8◦Cwith a total thermal requirement of 538 degree-days.Optimal development occurred at 29.6◦C. Stage dura-tion was progressively longer for later instars. However,in this study, the prepupal stage was not differentiatedfrom the sixth instar.

Ferreira (1995) determined the duration of eachinstar at a single temperature, 25◦C, while Viswanath(1976) and Gold et al. (1999c) measured instar dura-tion under ambient conditions in India (in differentmonths with temperatures not reported) and Uganda(20–27◦C), respectively. Combining the results of thesethree studies suggest that the relative proportion of timespent in each instar was: first instar (10%); second instar(11%); third instar (13%); fourth instar (14%); fifthinstar (15%); >sixth instar and prepupa (37%).

Under ambient temperatures in tropical environ-ments, the larval stage has been variously reportedas 2–3 weeks (Moznette 1920; Seshu Reddy et al.1998), 2–6 weeks (Simmonds 1966; Gordon & Ordish1966; Anonymous 1989), 3 weeks (de Villiers 1973;Waterhouse & Norris 1987), 3–5 weeks (Castrillon2000), 4 weeks (Gomes 1985), 5 weeks (Vittayaruket al. 1994), 3–6 weeks (Godonou 1999), 4–5 weeks(Carnero et al. 2002), 4–7 weeks (Longoria 1972),4–8 weeks (Trejo 1969; Mesquita & Caldas 1986),5–8 weeks (Padmanaban et al. 2001), 6 weeks (Ingles &Rodriguez 1989), 6–7 weeks (Cuille 1950), 7–8 weeks(Ferreira 1995), 6–9 weeks (Castrillon 1991; Aranzazuet al. 2000, 2001), 7–10 weeks (Padmanaban et al.2001), 7–17 weeks (Montellano 1954), 11–13 weeks(Velasco 1975). In Australia, the larval period canlast 23 weeks during winter months (Froggatt 1924).Mesquita et al. (1984) and Mesquita & Caldas (1986)determined the prepupal stage to be about 3 days, whileGold et al. (1999c) found it to be 4 days.

High variability in larval stage duration has evenbeen observed at the same locality, suggesting the influ-ence of food source and rearing methods on develop-mental rates. At one site near Kampala, Gold et al.(1999c) found the total larval and prepupal period

averaged 25 days. Rearing the larvae on thin corm slices(precluding formation of galleries) extended this to36 days. In contrast, Bakyalire (1992), working 20 kmaway and at a similar elevation, found the larval (andprepupal) period to range from 41 days (laboratory) to51 days (screenhouse).

The duration of the larval stage may be affected byclimate, food source, weevil density, and plant stage(Arleu & Neto 1984; Mesquita et al. 1984; Mesquita &Caldas 1986). For example, Mesquita & Caldas (1986)found the larval period was shorter on younger plants.Our calculations on their presented data for differ-ent clones show a mean larval stage of 29 days onyoung plants, 44 days on flowered plants and 52 dayson crop residues. The larval period also ranged from35 to 44 days when different clones were used ashosts.

d. Survivorship or survivalLarval survivorship or survival in the laboratory maybe reduced by handling and/or deterioration of the foodsource (Mestre 1997). However, Mesquita & Caldas(1986) successfully reared 87% of larvae to the pupalstage on young plants, 67% on flowered plants and 50%on crop residues. In addition, the developmental periodwas longer and pupal weights lowest on residues,suggesting this to be an inferior quality food. Traoreet al. (1996) successfully reared 50% of the larvaetested, with greatest mortality occurring in the first twoinstars. This may have reflected disturbance during thehandling of small larvae.

Under field conditions, mortality in the egg and first-instar stages appears to be high. Abera (1997) found6–12 times as many eggs as mid- to late-instar lar-vae during dissections of banana mats. Concurrent withlow oviposition rates, this may contribute to the slowpopulation build-up and greater importance of bananaweevil in ratoon crops (Mitchell 1980; Lescot 1988;Rukazambuga 1996).

3. Pupal stage

Pupation is in a bare chamber excavated by the larvaenear the corm surface of the host plant (Vilardebo 1960;Longoria 1972). Godonou (1999) found all pupae to bein the corm and most to be >5 cm below the collar.

In Benin, the pupal stage ranged from 5.5 days at30◦C to 23.0 days at 16◦C, with 10–19% mortality(Traore et al. 1996). The developmental threshold was10.1◦C with a thermal requirement of 121 degree-days.Optimal development occurred at 33.3◦C.

92 C.S. Gold et al.

Ferreira (1995) estimated the duration of the pupalstage at 25◦C to be 8.3 days. Under ambient tropi-cal conditions, the pupal stage has been most com-monly reported at 6–8 days (Froggatt 1925; Vilardebo1960; Haarer 1964; Saraiva 1964; Longoria 1972;de Villiers 1973; Viswanath 1976; Mesquita & Alves1983; Mesquita et al. 1984; Ingles & Rodriguez1989; Pinese & Piper 1994; Seshu Reddy et al. 1998;Godonou 1999; Castrillon 2000; Carnero et al. 2002)although stage durations of 6–12 days (Aranzazu et al.2001), 9 days (Padmanaban et al. 2001) and 11–13 dayshave also been reported (Montellano 1954; Gomes1985).

VII. Distribution and Population Dynamics

1. Population density and severity of attack

Densities of adult banana weevils in banana standshave been estimated by mark and recapture methods.In Cameroon, Delattre (1980) estimated weevil densi-ties in two fields to be 15,600/ha and 2600/ha, respec-tively; density within the fields varied from 1.1 to37.4 weevils/m2. Assuming a plant spacing of 3×3 m2,this would represent a range of 10–337 adults per mat.

In Mubende district, Uganda, Gold & Bagabe(1997) estimated weevil density at 900 adults/ha(3.4/mat) in a stand of cooking banana (cv Kibuzi,AAA-EA) and 2100/ha (4.8/mat) in an adjacent standof brewing bananas (Kisbui, AB, Ney Poovan sub-group), while Rukazambuga (1996) found a densityof 19,000 adults/ha in a banana stand (cv Atwalira,AAA-EA) in Mpigi District. In Uganda surveys, highvariability in weevil numbers and damage were alsofound among farms within-sites (Table 2). For exam-ple, population estimates ranged from 850 weevils/ha(1.5 weevils/mat) to 149,000 (240 weevils/mat) (Goldet al. 1997; Tinzaara & Gold unpubl. data).

Table 2. Estimated banana weevil population densities perhectare at selected sites in Uganda (N = 50–60 farms per site)

Site Range Median weevilnumber per matLow High Median

Ntungamo1 1600 149,000 9300 15Ntungamo2 2200 44,900 7900 13Mbarara 850 15,000 3500 6Masaka 2000 41,000 10,000 17

1Survey conducted in 1996.2Survey conducted in 1998.Sources: Gold et al. (1997), Masanza et al. (unpubl. data), Goldet al. (unpubl. data).

This within-site variability suggests that manage-ment plays an important role in regulating weevil pop-ulations. Weevil pressure has been widely believed tobe associated with low levels of management, stressedplants, bad drainage, acid or low fertility soils, weedyfields, inadequate sanitation, extended droughts, andnematode infestations (Froggatt 1925; Veitch 1929;Wallace 1938; Harris 1947; Gordon & Ordish 1966;Ostmark 1974; Loebel 1975; Yaringano & van der Meer1975; Firman 1970; Ambrose 1984; Van den Enden &Garcia 1984; Mesquita & Caldas 1986; Sebasigari &Stover 1988; Allen 1989; Boscan de Martinez & Godoy1989; Sikora et al. 1989; Bakyalire 1992; Treverrowet al. 1992; Speijer et al. 1993; Garcia et al. 1994;Pinese & Piper 1994; Stanton 1994; Gowen 1995;Sponagel et al. 1995; Schill et al. 1997). In centralUganda, farmers associated increasing banana weevilpressure with reduced management, and cited this asan important factor in the decline and disappearanceof highland cooking banana in the region (Gold et al.1999b).

In the Ntungamo survey, however, characterisationof study farms failed to reveal any important rela-tionships between management practices and weevilseverity (Gold et al. 1997). Additionally, Rukazambuga(1996) and Rukazambuga et al. (2002) found sim-ilar levels of weevil attack in stressed (intensivelyintercropped) and vigorous (mulched) banana systems(see below). Within banana stands, neither Treverrow(1993) nor Rukazambuga (1996) found any evidenceto suggest that smaller or thinner plants were predis-posed to greater levels of weevil attack than occurredin healthier plants. Moreover, weevil adult populationsmay be as much as 2.5 times higher in mulched thanin unmulched plots because of more favourable soilmoisture conditions (Price 1994; Rukazambuga et al.2002).

In Ghana, weevil damage in plantain stands wasoften low in spite of susceptible germplasm, favourabletemperatures and low management levels (Schill 1996).Shifting agricultural systems predominated with mostplantain abandoned after two crop cycles. As such,short plantation life may preclude adequate time toallow weevil populations to build up to damaging lev-els. Weevil problems became increasingly evident inplantain stands maintained beyond two cycles (Schill1996).

In Uganda, weevil damage was very variableacross survey sites (Gold et al. 1993, 1994a, 1999b).Damage was especially severe throughout much ofthe central zone where banana stands received limited


management attention and banana was in rapid decline.In these areas, banana weevil damage averaged 10% ofsurface area in the cortex and in the central cylinder(as measured in cross cut sections, Gold et al. 1994b).Elsewhere, however, it was difficult to discern whysome sites had higher levels of weevil damage thanothers.

2. Population build-up

Banana weevils have limited dispersal capacity andare most often disseminated through infested plantingmaterial. However, suckers used as planting materialusually support only a small number of weevil eggsand larvae (Gold et al. 1998a; Abera et al. 1999), soinitial populations are often low. With low fecundity,weevil build-up tends to be slow and several cycles maybe required before populations are fully established(Arleu & Neto 1984). In a survey of 52 sites in Ghana,Schill (1996) found cross section damage increasedfrom 2% in the plant crop, to 3% in the first ratoon and7% in the second ratoon. However, a large populationof banana weevils may quickly become established innew banana stands placed in or near previously infestedfields.

In Uganda, Rukazambuga (1996) monitored popula-tions with mark and recapture methods for 3 years fol-lowing release of 8750 weevils (mean of 9 weevils/mat)in a field trial 9 months after planting and 3 monthsbefore flowering. The released weevils had been col-lected from traps in a farmer’s field and were of uncer-tain age distribution. The population initially declinedby 44% to 4904 weevils at 8 months after release, thenrecovered to 9632 weevils at 12 months after releaseand peaked at 19709 weevils 32 months after release(a 4.5-fold increase in 24 months). Median cross sec-tion damage levels increased from 7% in the first ratoonto 10% in the second ratoon and >20% in the thirdratoon.

With existing estimates of weevil longevity, disper-sal capacity and fecundity, slow rates of populationbuild-up in young stands, and population fluctuations inestablished plantations suggest either high mortality inthe egg and larval stages and/or the influence of densitydependent factors on oviposition and survivourship.

In laboratory rearing experiments, Mesquita &Caldas (1986) reported 17–50% mortality in the larvalstage and 11–16% in the pupal stage, while Bakyalire &Ogenga-Latigo (1992) found 27–38% mortality in theegg stage, 9–23% in the larval stage and 0–4% in thepupal stage. Traore (1995) observed 43–53% mortality

in the larval stage, primarily in the first two instars. Lar-val performance will be influenced by clone and hostplant defence mechanisms. Under field conditions, sapor other antibiotic factors might cause additional mor-tality to the egg and first-instar larvae that might not berealised in the laboratory (Kiggundu 2000). It is unclearhow environmental factors might influence plantdefence and mortality levels of weevils immatures.

3. Adult populations and damage

Shillingford (1988) and Gowen (1995) have suggestedthat weevil adult density may not be closely related todamage levels. Gold et al. (1997) estimated weevil pop-ulations, percentage weevil damage and area (i.e. cm2)damaged (i.e. to account for between-farm differencesin plant size) in cross sections of recently harvestedplants on 50 farms in Ntungamo district, Uganda.Further analysis of the survey data sets revealed poorrelationships between weevil populations and percent-age damage (r = 0.11) and between weevil popula-tions and area damaged (r = 0.22) (Gold et al. unpubl.data).

These data have implications for control strategiesof banana weevil. Methods targeting adult weevilsmay require a considerable lag time before popula-tion reductions are translated into reduced damage andincreased yields.

4. Density dependence

Two factors suggest the possible existence of densitydependence effects on banana weevil oviposition inthe field. First, the rate of population build-up is oftenslower than expected, given the adult’s longevity andlimited dispersal capacity. Second, weevil damage maybe poorly related to population density suggesting thepossibility of reduced oviposition at higher densities(see below). However, Koppenhofer (1993a) felt thatdensities under field conditions would never reach thepoint of adversely affecting oviposition rates, in light ofthe weevil’s low reproductive potential and abundanceof host plants.

Density dependent effects on banana weevil oviposi-tion were first reported by Cuille (1950) who found thatgrouped weevils produced only 28–38% as many eggsas isolated individuals. Dissections of grouped femalesshowed they had fewer oocytes (of which some weremalformed) than ungrouped weevils.

Rukazambuga (1996) compared oviposition ratesover 4 weeks for densities of 2, 5, 10, 20 and 40 females

94 C.S. Gold et al.

(with equal numbers of males) on potted plants.Individual females at the highest two densi-ties produced 3.5–4.4 eggs/week compared to12.4–14.5 eggs/week at lower densities. As a result,groups of 40 females produced only 4.9 and 2.3 timesas many eggs as groups of 2 and 5 females, respectively.In another study (Gold et al. 2002a), offered pieces ofcorm to different densities of weevils maintained inbuckets. Mean oviposition rates per female at densi-ties of 10, 20, and 40 females per bucket were 29%,37%, and 53%, respectively, lower than that at a densityof five females. Nevertheless, total oviposition for thesame groups was 1.4, 2.5, and 3.7 times higher than thatof the five-female group. Oviposition rates were sim-ilar on corms changed daily and those changed every5 days, suggesting that the presence of eggs did notdeter further oviposition. Under field conditions, esti-mated oviposition per female declined with increasingweevil density with averages of 1.4 eggs/week at a den-sity of 5 females/mat, 0.8 eggs/week at 20 females/matand 0.5 eggs/week at 40 females/mat (Abera 1997).

Koppenhofer & Seshu Reddy (1994) suggest thathigh larval populations can result in cannibalism anddwarfism of adults. In contrast, Gold et al. (2002a)found larval survivourship was only slightly higher atlower densities of immatures following insertion ofdifferent densities of eggs or first-instar larvae intobanana corms. Gold et al. (2002a) concluded thatdensity dependent factors can influence ovipositionrates of individual weevils and survivourship of imma-tures but are likely to exert only modest influence inreducing banana weevil population growth under fieldconditions. More likely, high mortality of weevil imma-tures, independent of density, and/or higher rates ofadult mortality and emigration than previously pos-tulated contribute to the slow population build-up ofthis pest.

VIII. Semiochemicals

1. Host plant location

Olfactory cues may be utilised by banana weevilsin locating host plants, conspecifics and/or mates.Cuille (1950) was the first to propose and test‘chemotropisms’ of banana weevil. In a series of exper-iments using lures and olfactometers, both male andfemale weevils regularly oriented to corm material orwater, acetone and ether extracts of banana from dis-tances of up to 20 cm. The weevils were more attracted

to pseudostem than to corm material, but ovipositionwas greater on the latter. This suggested to Cuille thatolfactory cues were most important in host location,while chemoreception played the predominant role inhost acceptance. Cuille concluded that chemorecep-tion is probably more important than olfaction for asedentary insect like banana weevil.

Although Cuille (1950) and, later, Sumani (1997)found greater attraction to pseudostems than corm,traps containing corm material tend to be more effectivethan those made from pseudostems alone. Moreover, ithas long been recognised that freshly cut corms or suck-ers are especially attractive to the weevils (Sein 1934;Franzmann 1976; Jaramillo 1979; Treverrow 1994).

In still-air olfactometers, both male and femaleweevils oriented towards freshly chopped cormand pseudostem material, as well as to associatedPorapak-trapped volatiles from susceptible highlandcooking (AAA-EA) and resistant dessert (AB) bananas(Budenberg & Ndiege 1993; Budenberg et al. 1993b).Males and females also showed similar electroan-tennogram (EAG) responses to plant volatiles, suggest-ing that weevil orientation is to food sources ratherthan oviposition sites. Budenberg et al. (1993b) foundthe weevils demonstrated a somewhat lower responseto resistant AB than to susceptible AAA-EA clones,implying that antixenosis may contribute to host plantresistance. However, in other studies, no differenceswere found in attractivity to a wide range of clones(Pavis & Minost 1993; Abera 1997; Kiggundu 2000).Budenberg et al. (1993b) found that the major compo-nents of trapped volatiles did not elicit any responseindicating that minor components are responsible forweevil activity. Weevil orientation to plant volatilescould be demonstrated over only a few centimetres(Budenberg & Ndiege 1993; Budenberg et al. 1993b;Braimah 1997).

Jones (1968) reported that banana weevils appearto be highly attracted to certain lipophilic plant andAnnalose-11 volatiles from the cultivar Valery (AAA).More recently, Ndiege et al. (1991) and Lemaire (1996)identified mono- and sesquiterpenes as major compo-nents of volatiles from pseudostems and attractive toweevils. Ndiege et al. (1996a) found 1,8 cineole insusceptible or tolerant bananas and absent in resistantclones.

Using olfactometers and tests in laboratory arenas,Braimah (1997) and Braimah & van Emden(1999) found weevils were more attracted to deadbanana leaves that had undergone natural senes-cence on the plant than to the corm or pseudostem


(i.e. their oviposition sites). Surprisingly, the weevilswere more attracted to dead yam leaves and equallyattracted to hay as to banana leaves. They then tested theindividual chemical components that had been identi-fied (and previously tested collectively) by Ndiege et al.(1991, 1996a) as attractants, repellents or arrestants insearching behaviour of weevil. None of the chemicalsconstituents tested were more attractive than a distilledwater control, while some acted as repellents. Thesefindings agree with those of Budenberg et al. (1993b)and suggest that unidentified minor elements, non-volatiles factors, and/or different proportions of volatilecomponents are required to attract banana weevils.Braimah & van Emden (1999) concluded that bananaweevil attraction to host plants is a stepwise-mediatedprocess, by which senescing leaves draw weevils to thevicinity of the host plant.

In other work on volatiles using choice chambers,Padmanaban et al. (2001) found some attraction toleaf sheaths, but that most weevils were attracted tocorm pieces or leaf sheaths treated with corm extracts.Castrillon (2000) reported the weevil to be attracted tovolatiles from leaf veins and pseudostems, while Cerdaet al. (1995) and Reyes-Rivera (2000) found weevilsequally attracted to corms and pseudostems.

2. Pheromones and aggregation response

Cuille (1950) observed an aggregation response ofbanana weevils that he attributed to both olfactoryand tactile cues. He reported that olfactory cuesbrought weevils of the different sexes together, whilethigmotaxis stabilised aggregations.

Budenberg et al. (1993a) tested the presence ofpheromones by extracting volatiles from the headspaces above males and females, and by surface washesand extracts from the hindgut of each sex. These wereused in behavioural bioassays and EAG recordings.Both sexes were attracted to and made longer visitsto live males and male volatiles. Both sexes also hadEAG responses to male volatiles and hindgut extracts,while there was no response by either sex to femalevolatiles, washes or extracts. These results suggestedthat the males release an aggregation pheromone viathe hindgut that is attractive to both males and females.

Braimah (1997) confirmed the presence of a maleaggregation pheromone in male frass and hind parts.He concluded that the males were first attracted to theproximity of banana mats by dead leaves, followed byshorter range attraction to the mats themselves and,once arriving at the mat, produced a male aggregation

pheromone attractive to both sexes. In olfactometers,weevils were more attracted to corm pieces which hadpreviously supported male and female feeding than tocontrols or corms fed only on by females. Viana (1992)also suggested that chemical cues from the host plantwere needed to stimulate pheromone production.

Following on the results of Budenberg et al. (1993a),Beauhaire et al. (1995) identified and synthesised amale-produced banana weevil aggregation pheromone.Six male-specific compounds were isolated of which80% was comprised by a single compound (C11H20O2)with a formula as (1S∗, 3R∗, 5R∗, 7S∗) 2,8-dioxa,3,5,7-trimethyl bicylo (3,2,1) octane 1d. This com-pound, named sordidin, is related to known ketalpheromones produced by scolytids. Mori et al. (1996)identified the natural configuration of sordidin, whileJayaraman et al. (1997) synthesised four isomers(exo-B, endo-B, endo-A, exo-A) of sordidin. All fouroccur naturally in ratio of 1 : 4 : 4 : 44. Exo-B andexo-A evoked similar levels of antennal response inspite of their proportional differences. Ndiege et al.(1996b) and Jaramayan et al. (1997) found that a mix-ture of isomers were more attractive than single iso-mers. Wardrop & Forslund (2002) have also provideda synthesis for sordidin.

The use of pheromones and attractive plant volatiles(i.e. kairomones) as a means of controlling bananaweevil through mass trapping or in baits for delivery ofBeauveria bassiana was first proposed by Budenberget al. (1993a) and Kaaya et al. (1993). Jayaraman et al.(1997) also suggested semiochemical-enhanced masstrapping could overcome the low reproductive capac-ity of the insect and lead to successful control. Theyalso suggested that pheromones could be used in thedelivery of entomopathogenic nematodes. ChemticaInternational, Inc. in Costa Rica has begun commercialproduction of banana weevil pheromones (enhancedwith kairomones) in a product called Cosmolure+(Oehlschlager pers. comm.) which is being tested inLatin America, India and Africa.

IX. Sampling Methods

Accurate assessment of banana weevil populationlevels and damage are necessary to understand theweevil’s pest status, for screening germplasm and asa prerequisite in evaluating the impact of any inter-vention strategies. Sampling of banana weevil is diffi-cult and there are no agreed upon sampling protocols.The seclusive behaviour of the adult weevils and the

96 C.S. Gold et al.

difficulties in measuring larval damage in the interiorof the corm (much of it below ground level) in a peren-nial crop has resulted in a multitude of scoring andevaluation systems. These include monitoring adultnumbers through trapping; measuring damage to thecorm periphery; and measuring internal corm damage(e.g. through cross and longitudinal sections). Some ofthese methods are subjective, making it hard to interpretresults or to compare studies conducted by differentresearchers.

Damage evaluations require destructive samplingand are most often taken on recently harvested or deadplants. However, the weevil is an indirect pest andit is not clear what types of damage have the great-est impact on yield. In evaluating sampling methods,researchers should consider (1) ease of data collection;(2) precision; (3) accuracy of population or damageestimates; (4) degree of subjectivity and ability forother researchers to interpret the data (5) reflection ofpest status (Gold et al. 1994b).

1. Trapping

Trapping of adults as a means of assessing bananaweevil population levels has been employed since1912 (Knowles & Jepson 1912; Froggatt 1928; Cuille1950) and continues to be widely used (Arleu 1982;Bujulu et al. 1983; Allen 1989; Anonymous 1989;Ogenga-Latigo & Bakyalire 1993; Mestre & Rhino1997). Most commonly traps are made out of cropresidues (i.e. recently harvested corms and pseuod-stems). A variety of trap types, including disk onstump, corm disk, split pseudostem, sandwich, wedge,and canoe, have been described by Castrillon (1989,1991) and Aranzazu et al. (2000). Trap catches of upto 50 weevils have been reported (Suplicy Filho &Sampaio 1982).

Traps utilising corm material are generally moreattractive to banana weevils than those made from pseu-dostems (Edwards 1925; Hord & Flippin 1956; Saraiva1964; Martinez 1971). Yaringano & van der Meer(1975), Cardenas & Arango (1986) and Moreira et al.(1986) captured four to ten times as many weevils incorm-based traps (e.g. disk on stump) than in pseu-dostem traps. Most recommendations, however, con-cern the use of pseudostem traps for which material ismore readily available and can be placed anywhere.Traps made from the lower part of the pseudostem(e.g. 30–70 cm above the collar) may be more attrac-tive to weevils than those from higher on the plant(Mestre & Rhino 1997). Adding crushed corm to

pseudostem traps can triple weevil captures (Bakyalire1992).

Recommended trap densities for monitoring weevilpopulations range from ≤10/ha (de Villiers 1973;Crooker 1979; Castrillon 2000), 20–30/ha (Moreira1979; Suplicy Filho & Sampaio 1982; Arleu 1983;Castrillon 1991; Sponagel et al. 1995) to 40–60/ha(Allen 1979; Arleu 1982; Peasley & Treverrow 1986;Treverrow et al. 1992). Pest assessment based ontrap catches has operated on the assumption thatthe numbers captured are proportional to the pop-ulation size (Minost 1992). Action thresholds havebeen suggested at >1 weevil/trap (McNutt 1974; Allen1989), 2 weevils/trap (Mitchell 1978; Moreira 1979;Treverrow et al. 1992; Pinese & Piper 1994 (southeastQueensland)), 3 weevils/trap (Cuille & Vilardebo1963), 4–5 weevils/trap (Pullen 1973; Arleu 1983;Boscan de Martinez & Godoy 1989; Pinese & Piper1994 (north Queensland); Treverrow 1993; BatistaFilho et al. 1995b; Smith 1995; Aranzazu et al.2000), >6/weevils/trap (Stanton 1994) and (in CentralAmerica) >15 weevils/trap (Roberts 1955; Vilardebo1960; Pullen 1973; Anonymous 1989; Sponagel et al.1995). However, few studies have related trap catchesto damage or to yield loss and there appears to be littlejustification for any of these recommendations.

The use of trap catches as a means of assess-ing banana weevil populations has been criticised byVilardebo (1960, 1973). With long adult life (estimatedat 10–18 months) and a slow rate of increase, weevilpopulations should be relatively constant in establishedstands. However, trap catches can vary by a factor of3 or more within-sites during a single year. This sug-gests that the number of weevils coming to a trap areinfluenced by trap quality, humidity, temperature andany other factors influencing the behaviour of insect.For example, Bakyalire (1992) found that pseudostemtrap catches were affected by type of trap material(clone, age), size, location, number, soil moisture, andthe elapsed time between setting and checking traps.

Trap catches may also be a poor indicator of eco-nomic status if a high population estimate reflectsattack of residues rather than maturing plants. Gold &Bagabe (1997) collected twice as many weevils fromtraps in a stand of resistant beer banana (Kisubi, AB)than in an adjacent stand of susceptible cooking banana(Kibuzi, AAA-EA). Traps also provide no informationon which clones within a field may be most prone toattack. Weevils may also have more effect on plantswith smaller corms. For example, Vilardebo (1960)suggests that a given number of weevils cause 5 times


more damage on Gros Michel than on Cavendish dueto differences in timing of attack and corm size.

The use of different clones serving for trap materialcan greatly influence weevil captures, although resultshave not been consistent. Martinez (1971) reported thattraps from Giant Cavendish collected more weevils(4.0/trap) than from Dwarf Cavendish (0.8/trap) andthat traps from harvested pseudostems caught moreweevils (4.9) than traps from young pseudostems(2.7). Haddad et al. (1979) found traps from AABclones more attractive than from AAA clones. InUganda, traps made from Gros Michel attracted moreweevils (2.8 adults/trap) than traps made from high-land cooking banana, plantain, Ndiizi, or Kayinja(1.4–1.8/trap) (Bakyalire 1992), while in Kenya, trapsfrom highland cooking banana caught 2–12 times asmany weevils as traps from Gros Michel (K.V. SeshuReddy pers. comm). Gold & Bagabe (1997) collectedmore than twice as many weevils in traps made fromKisubi than from Kibuzi. Sumani (1997) reported trapcatches on Lusumba (AAA-EA) and Ndizi (AB) tobe more than twice those on Mbidde (AAA-EA) andNakyetengu (AAA-EA).

Trap placement can also influence trap captures.Traps placed against banana mats caught four times asmany weevils as those placed >10 cm away (Bakyalire1992). Traps next to older plants and stumps caught4 times as many weevils than those proximal to youngerplants.

Recommended trap sizes have ranged from25–30 cm (Schmidt 1965; de Villiers 1973; Bujulu et al.1983; Masanza 1995), 40 cm (Batista Filho et al. 1990)to 50–60 cm (Cuille & Vilardebo 1963; Simmonds1966; Arleu 1982; Suplicy Filho & Sampaio 1982). Inone study, larger traps caught more weevils than smallertraps, but smaller traps were twice as effective per unitlength than larger traps (i.e. 8.1 weevils per 35-cm trap;2.7 weevils per 5-cm trap) (Bakyalire 1992).

Recommended collection time intervals followingplacement of traps also vary: 1–3 days (Hord & Flippin1956; Vilardebo 1960; Delattre 1980; Bakyalire 1992;Aranzazu et al. 2000), 4–8 days (Wallace 1938; Cuille1950; Mitchell 1978; Arleu 1982; Bujulu et al. 1983;Treverrow 1985) to 14 days or more (Cardenas &Arango 1986). Gold et al. (1997, 2001) used 3–5 dayintervals between placement of pseudostem traps andweevil collection, as traps often deteriorated after5 days (C. Gold pers. observ.).

In contrast, Mestre & Rhino (1997) found weevilcaptures peaked at 7 days after trap placementalthough numbers were statistically similar between

5 and 14 days. Weevils became increasingly concen-trated under the true stem, which remained moist, com-pared to under the leaf sheaths, which tended to dry outover time. Equal numbers of males and females wereattracted to fresh traps (i.e. at 2 and 4 days), but malesconstituted 63% and 57% of the weevils collected in7- and 14-day-old traps, respectively. Mestre & Rhino(1997) recommended collection of weevils 7–9 daysafter trap placement.

A number of authors have reported a poor relation-ship between weevil trap captures and corm damage(Haddad et al. 1979; Crooker 1979; Ogenga-Latigo &Bakyalire 1993). In one trial, trap catches had correla-tions of 0.27 with peripheral damage to the corm and0.05 and 0.28, respectively, with internal corm dam-age estimates taken in cross and longitudinal sections,respectively (Ogenga-Latigo & Bakyalire 1993).

In contrast, Mestre & Rhino (1997) found that dam-age closely followed trap catches in the third and fourthcycle of a trial. They concluded that trapping doesgive a good indicator of damage and recommendedthat this should be done at the time of plant flower-ing. It is unclear, however, how this should be done inestablished systems where flowering occurs through-out the year. Moreover, all indications do suggest thattrap captures are clearly influenced by climatic andother factors. For example, capture rates in one ofMestre and Rhino’s trials varied over time from 1.5to 5.0 weevils/trap, from 0.1 to 9.2 weevils/trap in asecond trial and from 0.6 to 8.5 weevils/trap in a third.

In a screening trial against banana weevil, Kiggundu(2000) also found no relationship between trap cap-tures at the base of the different clones and the damageoccurring in those clones. Similarly, Abera et al. (1999)trapped comparable numbers of weevils on three high-land cooking clones, two highland brewing clonesand Kayinja (ABB), yet each of the highland cloneshad much greater damage levels than Kayinja. BatistaFilho et al. (1992) found similar trap catches (threeweevils/trap) on the banana clones Nanica and Prata,even though Prata was found to be much more suscep-tible to the weevil. Ogenga-Latigo & Bakyalire (1993)found more weevils trapped on AB than AAA-EAclones but 4–8 times more damage on the latter.

The relationship between trap catches and weevilpopulations is likewise unclear. Reinecke (1976) foundfour times as many weevils in uprooted plants than intraps, while Cardenas & Arango (1987) suggest that5–7 times as many weevils are attracted to the bananamat as to traps. Due to the large variation in meth-ods employed and environmental conditions, trapping

98 C.S. Gold et al.

provides imprecise estimates of weevil populations thatare inconsistent, statistically unreliable, and difficultto compare and interpret (Ogenga-Latigo & Bakyalire1993).

Weevil populations can be more accurately esti-mated using standard mark and recapture methods(Delattre 1980; Price 1993; Gold & Bagabe 1997; Goldet al. 1997) including trapping, marking, releasing, andsubsequent re-trapping. Many adult weevils are inac-tive for extended periods (S. Lux pers. comm.), sug-gesting that a 1-week interval elapse between weevilrelease and re-trapping to allow for a thorough mixingwithin the field population. It is also important that anadequate number of traps be placed randomly withinthe banana stand.

Using mark and recapture methods, Price (1993) andRukazambuga (1996) found greater weevil density inmulched plots than in unmulched plots, even thoughtrap captures were higher in the latter treatment. Inother words, trap efficiency was lower in the mulchedsystems. These results highlight the weaknesses inusing simple weevil trap capture rates to estimateweevil populations or to serve as action thresholds.

2. Plant damage

Assessment of weevil damage to banana corms requiresdestructive sampling and is most often conducted onharvested plants. In some clones (e.g. Kisubi), there issubstantial attack of residues (Gold & Bagabe 1997)which will have no bearing on plant yield. Therefore,it is important that sampling be done as soon afterharvest as possible. All existing sampling methodsmeasure cumulative weevil attack throughout the lifeof the plant and cannot determine when this attackoccurred. Timing of attack must be inferred throughstudies on oviposition (Abera 1997) and larval per-formance (Mesquita & Caldas 1986) on different hostphenological stages.

a. Proportion of damaged plantsVilardebo (1960) first proposed estimating the percent-age of harvested plants with weevil damage in theperiphery as an alternative to trapping. He felt that aproportion greater than 10% justified treatment.

Mestre (1997) and Mestre & Rhino (1997) com-pared the percentage of plants attacked (grouped intofive classes) with an estimate of damage to the cormperiphery (modified from Vilardebo’s (1973) coeffi-cient of infestation (CI) and grouped into four classes).A strong curvilinear relationship and close correlation(r = 0.98) existed between the two sets of values.

Based on this, Mestre proposed that percentage ofplants attacked is a better indicator of pest status thanactual damage measurements because it requires lesswork. Heavy damage could be said to exist when morethan 50% of the plants have been attacked. However, inUganda surveys,>50% attack of plants was observed atall sites under 1600 masl, even though peripheral dam-age varied by a factor of >3 among sites (Gold et al.1994a, unpubl. data). Moreover, no comparisons weremade in the Mestre (1997) study between the percent-age of plants attacked and damage parameters (i.e. tothe central cylinder or cortex) that may have greaterimpact on yield (Ogenga-Latigo & Bakyalire 1993;Rukazambuga 1996).

b. Number of weevil tunnelsVilardebo (1960) offered that visual observations onlarval damage are a better indicator of weevil pest statusthan trap counts and recommended counting the num-ber of galleries exposed on the corm periphery. Romanet al. (1983) and Ogenga-Latigo & Bakyalire (1993)counted tunnel points on cross sections through thecorm, while Treverrow et al. (1992) estimated the num-ber of galleries >4 mm in diameter in 75% of the areaof a single cross section through the corm. Althoughtunnel points are sometimes clear (especially at lowlevels of damage), in many cases they are not, as theymay converge or be associated with rots.

c. Harvested plants – peripheral damageVilardebo (1973) was the first to propose using areaattacked as a means of estimating the degree of dam-age to the plant. He suggested that this was preferableto counting the number of galleries, which often con-tained a high degree of error. According to Vilardebo,weevil larvae prefer the upper cortical zone of the corm,while under conditions of heavy attack, the larvae mayspread first to the lower parts of the corm, then theinterior of the corm and, finally, to the pseudostem andfollowers. To measure this damage, Vilardebo (1973)developed the ‘coefficient of infestation’ (CI) whichentails paring the corm below the collar to a depth of1–2 cm. The CI then represents an estimate of the pro-portion of the corm surface (i.e. 0–100%) with weevilgalleries. Although this is a destructive technique, itcan often be done ‘in situ’ with minimal damage to thestability of the mat.

Vilardebo (1973) recommended a sample sizefor estimating weevil pest status of at least30 plants/ha/date. Vilardebo (1973) also provided rela-tionships between trap catches and damage scores andbetween damage scores and yield loss. However, it


is unclear how these relationships were determinedas no experimental designs, data sets or analyseswere described. Therefore, they must be regarded withcaution.

The CI has been widely used as a means of estimat-ing weevil damage (Crooker 1979; Haddad et al. 1979;Englberger & Toupu 1983; Kehe 1985; Lescot 1988;Fogain & Price 1994; Smith 1995). However, the CIscoring system was not clearly defined and remainshighly subjective, making interpretation of reportedscores difficult. A damage assessor must visualise agrid by which he can evaluate the area of the corm withweevil damage. This grid can be fine or coarse. As aresult, different scorers may assign a wide range ofvalues for similar levels of damage. Thus, it is unclearif the CI values reported by researchers in Australia,Cameroon, Cote d’Ivoire, Tonga, and Venezuela arecomparable.

To standardise the scoring system, Mitchell(1978, 1980) developed a ‘percentage coefficient ofinfestation’ (PCI). This method entails removal of10 cm of topsoil and exposure of a band (7 cm wideand 1

2-cm deep) at the widest point of the corm.

Presence/absence of peripheral weevil damage isrecorded in 10 sections, each covering 18◦ of the cormsurface with the PCI representing the number of sec-tions with weevil damage (i.e. 0–10). However, the PCItends to saturate quickly and may underestimate highlyclumped damage (Bridge & Gowen 1993; Gold et al.1994b). Gold et al. (1994a,b) employed a modified PCIby increasing the grid to 20 sections but found this pro-vided no greater resolution in damage scores (i.e. thecorrelation of the two PCI measurements was >0.9).

Given the weaknesses in both the CI and PCI meth-ods, Gold et al. (1994b) developed a scoring systemfor ‘peripheral damage’ that entailed paring the halfthe corm for 10 cm below the collar and estimatingthe percentage of surface area consumed by weevillarvae (i.e. taken up by galleries). Other measure-ments of peripheral damage include a coefficient ofhealth (Sampaio et al. 1982) and a mean CI (Mestre1997) (both derived from the CI scale), a coefficientof damage (Bujulu et al. 1983), and a weevil coeffi-cient of infestation (WCI) based on the length of weevilgalleries (Bosch et al. 1996).

d. Harvested plants – internal damageInternal damage within the corm is more likely to have agreater impact on banana growth and bunch filling thandamage to the corm periphery. Moreover, the abilityof the weevil to penetrate the corm may be cultivarrelated; as such, the CI and PCI serve as poor estimates

of internal corm damage (Ogenga-Latigo & Bakyalire1993; Gold et al. 1994b).

Crooker (1979) was the first to measure internalweevil damage to banana corms, which was done byexamining galleries in cross cuts (‘cut corm method’).Mesquita (1985) proposed taking a longitudinal cuton one-fourth of the corm circumference at its widestpoint, with scores ranging from 0% to 100%. In bothmethods, it is unclear if the scoring system was derivedfrom Vilardebo’s (1973) CI or represented the propor-tion of surface area consumed by weevil larvae (i.e. ingalleries). Lescot (1988) employed the CI for tangentialcuts in the corm.

Taylor (1991) proposed adapting a presence/absencescoring system for a grid applied to five sections(i.e. modified PCI) of a transverse cut made at thecollar of the banana corm. Smith (1995) suggestedscoring presence/absence of nine sections in a similarlongitudinal cut in the corm periphery to that employedby Mesquita (1985). Treverrow (1993) used cross sec-tions, divided the surface into quarters and sampled allbut the quarter facing the follower. Damage was ranked0–3 in the remaining quarters (giving a range of 0–9)on the basis of the number of gallery points.

The most detailed scoring of internal damage wascarried out by Gold et al. (1994a,b) and Rukazambuga(1996). Two cross cut sections were made (at the col-lar and 10 cm below the collar). In each section, theyestimated the percentage of surface area consumed byweevil larvae in the central cylinder and in the outercortex. In addition, the area of each section was esti-mated by taking the diameter of the central cylinder andof the corm. The central cylinder and cortex containroughly the same area (Rukazambuga 1996), allow-ing the derivation of total internal damage to the cortex(mean value of cylinder and cortex scores). These mea-surements also allowed for conversion of percentagearea into square centimetres damaged.

In most cases, measurements of internal damagerequire partially or totally digging out of the corm, withresulting weakening of the stability of the mat. All ofthe proposed sampling procedures measure damage tothe top 10 cm of the corm. However, damage may bemuch more severe in the lower third than the upperthird of the corm (Englberger & Toupu 1983; Price1994; Gold & Kagezi unpubl. data).

e. Harvested plants – damage scalesMany of the scoring systems are imprecise orsubjective. For example, the CI depends on the asses-sor’s interpretation of the Vilardebo scale, whilepresence/absence systems may saturate quickly and

100 C.S. Gold et al.

do not account for clumping of damage. Estimatesof the surface area taken up by weevil galleries aremore precise but remain subjective as the scorer mustdetermine what is weevil damage and then estimateproportions.

Therefore, a number of researchers have indepen-dently proposed or employed damage scales with4–6 categories ranging from no attack to very severe(Liceras et al. 1973; Bendicho & Gonzalez 1986;Van den Enden & Garcia 1984; Treverrow et al. 1992;Bridge & Gowen 1993; Price 1994; Pinese & Piper1994; Sumani 1997). These scores are useful for com-paring treatment effects within studies but may not bereplicable by different research teams.

X. Pest Status and Yield Loss

The banana weevil has been considered a major bioticconstraint in Africa (Persley & de Langhe 1987;INIBAP 1988a,b; Gold et al. 1993), Asia and the Pacific(Valmayor et al. 1994) and Latin America (Arleu &Neto 1984; Mesquita et al. 1984; Arroyave 1985; Penaet al. 1993; Schmitt 1993; Castrillon 2000). The bananaweevil has been implicated as an important causeof the decline and disappearance of highland banana(AAA-EA) in its traditional growing areas in EastAfrica (Gold et al. 1999b; Mbwana & Rukazambuga1999).

Yield losses to banana weevil have been associ-ated with sucker mortality, reduced bunch weights andshorter stand longevity. Ovipositing weevils areattracted to cut corms (e.g. detached suckers used asplanting material), and can be serious pests duringthe crop establishment phase. In newly planted fieldswith existing weevil populations and no alternative hoststages, plant mortality may be high (Ambrose 1984;Price 1994; Gowen 1995; McIntyre et al. 2002). A sin-gle larva can kill a sucker if it attacks the growing point(C. Gold pers. observ.).

In one trial in Uganda, 40% of newly planted, weevil-free (i.e. pared and hot water treated) highland cookingbanana suckers were killed by banana weevils remain-ing from an earlier trial (McIntyre 2002). Messiaenet al. (2000) found mortality of weevil-free plantainsuckers at 34% in control plots during the first 3 monthsafter planting. Afreh-Nuamah (1993) also found highmortality of suckers planted in clean fields, when plant-ing material had been obtained from a heavily infestedfield.

In fields with minor initial infestations of bananaweevils, population build-up is slow with problems

likely to become increasingly important in ratoon crops(Reinecke 1976; Arleu 1982; Kehe 1985; Allen 1989;Staver 1989; Afreh-Nuamah 1993; Rukazambuga et al.1998; Gold et al. 1998b). Reinecke (1976) suggestedthat yield reductions first appear in the second ratoon,while Allen (1989) reported weevil problems to berelatively minor for 5 or 6 cycles. In a trial with high-land cooking banana, Rukazambuga et al. (1998) esti-mated yield loss to increase 5-fold from the first to thirdratoons.

Nevertheless, the pest status of banana weevilremains controversial. The banana weevil has vari-ously described as being an ‘important’, ‘serious’ or‘major’ banana production constraint in a region tobeing of local, restricted or no importance (Table 3).In Colombia, for example, the banana weevil has beenreported as being of ‘restricted importance’ (Castrillon1987; Anonymous 1992), the ‘most important andwidely distributed pest’ on banana (Cardenas & Arango1987; Castrillon 1989) to being a ‘major economicpest’ (Londono et al. 1991). In the Philippines, Davide(1994) suggested the weevil was a major produc-tion constraint and second in importance to nema-todes on Cavendish plantations, while Dawl (1985)felt the weevil pest problems were restricted to a fewplantations.

Banana weevil pest status may reflect banana type,clone selection, ecological conditions and manage-ment systems. Bananas are grown from sea levelto >2000 masl, under a range of different rainfalland soil conditions and in production systems rang-ing from kitchen garden to large-scale commercialplantations. Weevil damage levels are likely to bevery different on Cavendish bananas (AAA) grownin commercial plantations than on highland cookingbanana (AAA-EA) or plantains (AAB) grown in small-holdings. Pest status within genome groups is alsoin dispute. For example, recommended action thresh-olds on Cavendish banana vary from 2 weevils/trapin Brazil (Moreira 1979) to 15–25 weevils/trap inCentral America (Anonymous 1989; Sponagel et al.1995).

Few studies have attempted to quantify yield loss tobanana weevil. On-farm assessments have mostly con-cerned evaluation of plant loss through toppling andsnapping which can be attributed to banana weevils.Some of these studies failed to partition damage effectsof weevils and nematodes and yield loss estimates are,thus, confounded. Additionally, on-station trials aremost appropriate if yield losses are followed for severalcrop cycles. Only limited information is available on


Table 3. Summary of literature reporting different pest status of the banana weevil C. sordidus and suggested mechanisms of yield loss

a. Pest status‘Major’ pest or constraintJardine (1924), Hargreaves (1940), Sen & Prasad (1953), Hall (1954), Vilardebo (1960), Gorenz (1963), Wolfenberger (1964),Nonveiller (1965), Singh (1970), Edge (1974), McNutt (1974), Suplicy Fo & Sampaio (1982), Arleu & Neto (1984), Van den Enden &Garcia (1984), Arroyave (1985), Kehe (1985), Prando et al. (1987), Stover & Simmonds (1987), Tezenas du Montcel (1987), Lescot(1988), Swennen et al. (1988), Sikora et al. (1989), Sarah (1990), Batista Filho et al. (1991), Londono et al. (1991), Taylor (1991),Minost (1992), Varela (1993), Gold et al. (1993, 1999b), Simon (1993), Davide (1994), Pone (1994), Price (1994), Mukandalaet al. (1994), Sarah (1994), Stanton (1994), Vittayaruk et al. (1994), Bosch et al. (1996), Cerda et al. (1995), Ahiekpor (1996), SeshuReddy et al. (1998), Nkakwa (1999), Aranzazu et al. (2000, 2001), Castrillon (2000), Carnero et al. (2002)

Regional importanceVilardebo (1984), Waterhouse (1993), Gowen (1995)

Local, restricted, or of no importanceOstmark (1974), Firman (1970), Stephens (1984), Dawl (1985), Kusumo & Sunaryono (1985), Sebasigari & Stover (1988), BoscanMartinez & Godoy (1989), Anonymous (1992), Schill et al. (1997), Sponagel et al. (1995)

b. Mechanisms of yield lossSucker lossMoznette (1920), Hargreaves (1940), Weddell (1945), Stover & Simmonds (1987), Boscan de Martinez & Godoy (1989), Pena &Duncan (1991), Afreh-Nuamah (1993), Pinese & Piper (1994), Price (1994), Ndege et al. (1995), McIntyre et al. (2000), Messiaen et al.(2000)

Root initiationMontellano (1954), Vilardebo (1960, 1977), Cuille & Vilardebo (1963), Singh (1970), Waterhouse & Norris (1987), PCAARD (1988),Allen (1989), Castrillon (1989, 1991), Treverrow (1985, 1993), Treverrow et al. (1992)

Root death, reduced root numberSwaine (1952), Montellano (1954), Nonveiller (1965), Singh (1970), Wright (1977), Castrillon (1989, 1991), Kehe (1985, 1988), Lescot(1988), Londono et al. (1991), Musabyimana (1999)

Nutrient uptake/transportMoznette (1920), Montellano (1954), Vilardebo (1960, 1977), Cuille & Vilardebo (1963), Nonveiller (1965), McNutt (1974), Medinaet al. (1975), Kehe (1985, 1988), Treverrow (1985, 1993), Waterhouse & Norris (1987), PCAARD (1988), Allen (1989), Boscan deMartinez & Godoy (1989), Castrillon (1989, 1991), Sponagel et al. (1995), Masso & Neyra (1997), Musabyimana (1999), McIntyreet al. (2000)

Leaf senescenceHarris (1947), Cuille (1950), Nonveiller (1965), PANS (1973), Medina et al. (1975), Hely et al. (1982), Jones (1986), Cardenas &Arango (1987), Prando et al. (1987), Waterhouse & Norris (1987), Allen (1989), Ingles & Rodriguez (1989), Londono et al. (1991),Pinese & Piper (1994), Ndege et al. (1995), Sponagel et al. (1995), Masso & Neyra (1997)

Reduced plant size/vigourFroggatt (1925), Harris (1947), Cuille (1950), Montellano (1954), Roberts (1958), Cuille & Vilardebo (1963), Shell (1967), Deang et al.(1969), Trejo (1969), Singh (1970), Medina et al. (1975), Reinecke (1976), Kehe (1985, 1988), Jones (1986), Prando et al. (1987),Boscan de Martinez & Godoy (1989), Sikora et al. (1989), Pinese & Piper (1994, Ndege et al. (1995), Seshu Reddy et al. (1995),Rukazambuga (1996), Masso & Neyra (1997), Rukazambuga et al. (1998), Ngode (1998), Musabyimana (1999)

Increased susceptibility to drought/diseaseHarris (1947), Ambrose (1984), Uzakah (1995)

Slower growth rateRoberts (1958), Franzmann (1976), Rukazambuga (1996), Rukazambuga et al. (1998), Gold et al. (1998b)

Secondary invasionMontellano (1954), Hord & Flippin (1959), Mesquita et al. (1984), PCAARD (1988), Londono et al. (1991), Sponagel et al. (1995),Musabyimana (1999), Godonou et al. (2000)

Castniomera humboldtiArroyave (1985), Castrillon (1989, 1991), Carballo & de Lopez (1994), Londono et al. (1991), Carnero et al. (2002)

Pseudomonas (Ralstonia) solanacearumVilardebo (1960, 1977), Trejo (1969), Arroyave (1985), Batista Filho et al. (1987), Castrillon (1989, 1991, 2000), Londonoet al. (1991), Carballo & de Lopez (1994), Aranzazu et al. (2000, 2001)

Fusarium oxysporumTrejo (1969), Suplicy Filho & Sampaio (1982), Arroyave (1985), Castrillon (1989, 1991), Londono et al. (1991)

RotsHord & Flippin (1959), McNutt (1974), Mesquita et al. (1984), Kehe (1985, 1988), Williams et al. (1986), Waterhouse & Norris (1987),Shillingford (1988), Castrillon (1989, 1991), Treverrow et al. (1992)


Table 3. (Continued)

Plant lossHargreaves (1940), Simmonds & Simmonds (1953), Vilardebo (1960, 1977), Trejo (1969), Ambrose (1984), Prando et al. (1987),Castrillon (1989, 1991), Deang et al. (1989), Londono et al. (1991), Carballo & de Lopez (1994), Sponagel et al. (1995), Rukazambuga(1996), Masso & Neyra (1997), Rukazambuga et al. (1998), Aranzazu et al. (2000, 2001), McIntyre et al. (2000), Messiaen et al.(2000), Carnero et al. (2002), Gold et al. (unpubl. data)

SnappingHord & Flippin (1959), de Villiers (1980), Treverrow (1985, 1993), Shillingford (1988), Sikora et al. (1989), Taylor (1991), Stanton(1994), Gowen (1995), Sponagel et al. (1995), Bosch et al. (1996), Rukazambuga (1996)

TopplingSwaine (1952), Sen & Prasad (1953), Simmonds & Simmonds (1953), Braithwaite (1958), Roberts (1958), Vilardebo (1960, 1977),Cuille & Vilardebo (1963), Nonveiller (1965), Trejo (1969), PANS (1973), McNutt (1974), Medina et al. (1975), Jaramillo (1979),Roman et al. (1983), Ambrose (1984), Mesquita et al. (1984), Kehe (1985, 1988), Jones (1986), Williams et al. (1986), Prando et al.(1987), Waterhouse & Norris (1987), Stover & Simmonds (1987), Lescot (1988), Shillingford (1988), Ingles & Rodriguez (1989),Deang et al. (1989), Sikora et al. (1989), Marcelino & Quintero (1991), Pena & Duncan (1991), Pinese & Piper (1994), Ndege et al.(1995), Sponagel et al. (1995), Uzakah (1995), Rukazambuga (1996), Masso & Neyra (1997), Musabyimana (1999), Godonou et al.(2000), Ysenbrandt et al. (2000)

Reduced bunch size/number1

Sen & Prasad (1953), Roberts (1958), Vilardebo (1960, 1973, 1977), Braithwaite (1967), Shell (1967), PANS (1973), Franzmann(1976), Reinecke (1976), Roman et al. (1983), Kehe (1985, 1988), Jones (1986), Cardenas & Arango (1987), Shillingford (1988),Castrillon (1989, 1991), Sikora et al. (1989), Carballo & de Lopez (1994), Ndege et al. (1995), Sponagel et al. (1995), Rukazambuga(1996), Masso & Neyra (1997), Alpizar et al. (1998), Gold et al. (1998b), Rukazambuga et al. (1998), Ngode (1998), Aranzazu et al.(2000, 2001), Ysenbrandt et al. (2000), Carnero et al. (2002)

Reduced plantation lifeGold et al. (1993, 1999b), Masso & Neyra (1997), Castrillon (2000)

Reduced number of suckersFroggatt (1925), Sen & Prasad (1953), Franzmann (1976), Reinecke (1976), Roman et al. (1983), Ambrose (1984), Arroyave (1985),Jones (1986), Prando et al. (1987), Castrillon (1989, 1991), Uronu (1992), Ndege et al. (1995), Seshu Reddy et al. (1995), Sponagelet al. (1995)

Reduced vigour of suckersFroggatt (1925), Roberts (1958), Ambrose (1984), Arroyave (1985), Jones (1986), Uronu (1992), Pinese & Piper (1994), Ndege et al.(1995)

Water suckersBraithwaite (1958), Uronu (1992), Sponagel et al. (1995), Bosch et al. (1996)

1Implicit in most reports but not always explicitly stated.

the physiological basis of yield loss in banana resultingfrom banana weevil attack.

1. Weevil damage mechanisms

Various authors have proposed that banana weevilattack can kill young suckers, interfere with root ini-tiation in growing plants, kill existing roots, reducenutrient uptake and transport, lead to premature leafsenescence, reduce plant size, vigour and tolerance ofother stresses, retard maturation rates, increase sus-ceptibility to other pests and diseases, lead to plantloss through death before producing a bunch, top-pling and snapping, reduce the number of harvestedbunches and bunch weights, and affect the numberand vigour of followers (Table 3). Many of thesereports appear to be based on conjecture or fieldobservations without supporting data. Some may also

confound the combined effects of weevil and nematodeattack.

In established plants, timing and location of weevilattack within the corm may have distinct effects onplant development and yield. Weevil larvae may movefrom the mother plant and can kill young suckers.However, ovipositing females prefer flowered plants(Abera 1997) and this may explain, in part, why larvaldamage has more impact on bunch weight than on plantsize and maturation time (Rukazambuga et al. 1998).

Larvae feeding on the corm periphery (as measuredby the CI and PCI) can cut through root points of attach-ment (Montellano 1954; Singh 1970; Londono et al.1991). In Uganda, the number of detached roots inhighland banana was proportional to the area of cormsurface attacked (A. Abera & C. Gold unpubl. data).

By contrast, internal damage has been hypothesisedto affect root initiation, nutrient transport, and stem


growth, while more peripheral damage may detachroots or adversely affect root development (Taylor1991; Gold et al. 1994b). Larval feeding can inter-fere with root initiation and development (Lescot 1988;Boscan Martinez & Godoy 1989; Castrillon 1991;Treverrow et al. 1992) resulting in production offewer roots or premature root death (Swaine 1952;Montellano 1954; Vilardebo 1960; Cuille & Vilardebo1963; Nonveiller 1965; Cerda et al. 1995). Damageto the central cylinder has the greatest effect on thevascular system and may stunt stem growth while dam-age to the cortex may adversely affect root developmentand lead to snapping and toppling (Taylor 1991; Goldet al. 1994b). In Uganda, Gold et al. (1994a) found dif-ferences among genome groups and clones in degreeof penetration into the central cylinder: Plantains hadextensive damage throughout the corm, while in GrosMichel (AAA), Ndizi (AB), and Ney Poovan (AB)larval feeding was largely restricted to the cortex. Thecombined effects of weevil damage to the root andvascular system have been reported to disrupt nutri-ent uptake and transport (Moznette 1920; Montellano1954; Vilardebo 1960; Cuille & Vilardebo 1963;Nonveiller 1965; McNutt 1974; Castrillon 1991; Gold1998a) resulting in premature leaf senescence, stunted,weak plants and reduced bunch filling (i.e. loweryields).

Montellano (1954) suggested that larvae do moredamage when tunnelling near attachment points ofroots in the central cylinder or when near vascular bun-dles in cortex. Rukazambuga (1996) also concludedthat damage to the central cylinder had more effect onplant growth and yield than damage to the cortex. Incontrast, Boscan de Martinez & Godoy (1989) postu-lated that larval feeding on the corm surface proximalto the roots caused the greatest damage by impedingnutrient uptake.

Nevertheless, there are few available data quantify-ing the effect of weevil damage on nutrient uptake andfew controlled studies documenting its effect on leaflife, growth and yield. In a series of trials in Uganda,McIntyre et al. (2002, unpubl. data) and Ssali et al.(unpubl. data) concluded that weevil infestations pre-vented plants from taking advantage of nutrient amend-ments to the soil. Hassan (1977) observed that damageprevents water from reaching the leaves causing themto yellow and wither, while, according to Jones (1986)attacked plants have dull, flaccid leaves.

Rukazambuga (1996) and Rukazambuga et al.(1998) demonstrated that the effects of weevil damageon yield was influenced by the levels of weevil damage

to the mat in earlier cycles; i.e. weevil attack influencedthe vigour of followers. Gold et al. (unpubl. data) found>35% of highland banana mats died out in 5 years inplots infested with weevils, compared to 2% mat lossin controls. This suggests that the weevil can severelyreduce stand life. Farmers in central Uganda reportedthat the weevil had contributed to reductions in standlife from >30 years to 4 years (Gold et al. 1999b),while in Colombia the weevil can reduce stand life to2–3 crop cycles (Castrillon 2000).

2. Toppling and snapping

Toppling is often attributed to plant parasitic nematodes(e.g. Radopholus similis, Pratylenchus spp.) that attackthe root system, thereby reducing anchorage (Gowen1995). However, it is likely that weevil damage reducesroot number and can contribute to toppling. For exam-ple, in Ugandan field trials, Rukazambuga (1996) foundextensive toppling in mats with high levels of weevildamage and negligible levels of nematodes and rootnecrosis. Similar observations have also been made onfarms in Bukoba District, Tanzania (N. Rukazambugapers. comm.). Snapping (i.e. breaking of the corm) mayalso occur on plants with severe weevil damage (Gowen1995). Most often, toppling and snapping occur onolder plants bearing the weight of a mature bunch.Nevertheless, toppling and snapping of maiden suckersmay also occur, especially following strong winds.

3. Plant stress and banana weevil attack

Stressed plants have been reported as being more attrac-tive or suitable for banana weevils than vigorously-growing plants (Wallace 1938; Nonveiller 1965;Ambrose 1984; Mesquita et al. 1984; Vilardebo1984; Allen 1989; Speijer et al. 1993; Mestre 1997).For example, banana weevil damage may behigher on nematode infested plants (Speijer et al.1993; Treverrow 1993) although data are limited.However, Rukazambuga (1996) found that initial attack(i.e. following release of weevils in an establishedstand) was independent of plant size and vigour.

Stressed plants have also been reported as dis-proportionately affected by banana weevil damage,while vigorous plants have been reported to be moretolerant to weevil damage (Froggatt 1925; Harris1947; Cuille & Vilardebo 1963; Pinese & Piper 1994;Sponagel et al. 1995). Rukazambuga (1996) found thatbananas in a mulched stand had a higher damage thresh-old than intensively intercropped bananas, but above


that threshold, percentage reductions in bunch weightwere similar.

Larval feeding may provide entry point fordisease agents such as Pseudomonas (Ralstonia)solanacearum (Moko disease) and other organismscausing rots (Table 3). Reports on the weevil vector-ing or providing entry points for Fusarium oxysporumSchlecht f.sp. cubense (E.F. Smith) Snyder & Hansen,a fungal pathogen causing Fusarium wilt (Suplicy Fo &Sampaio 1982; Castrillon 1991; Londono et al. 1991),have not been substantiated and may be in error.In Latin America, the moth borer, Castniomerahumboldti, has been reported as utilising weevil gal-leries to gain entrance into banana corms (Arroyave1985; Castrillon 1991; Carballo & de Lopez 1994).Ants (e.g. Ondotomachus troglodytes Sanschi) occa-sionally nest in and enlarge weevil galleries in livingplants although more often they enter crop residues(Gold pers. observ.). It has also been suggestedthat weevil damaged plants are more susceptible todrought and disease stresses (Harris 1947; Ambrose1984).

4. Pest status in Asia

The banana weevil is presumed to have evolved insoutheast Asia (especially the Indo-Malay region)from which it has spread to all of the world’smajor banana-growing regions (Zimmerman 1968b;Neuenschwander 1988; Waterhouse 1993). Cuille(1950) suggested that the banana weevil was not apest in Indonesia due to the prevalence of resistantclones and climate. However, data on the weevil’spest status (e.g. distribution, incidence, severity, yieldloss) in its area of origin are sparse (Waterhouse1993; Valmayor et al. 1994). Scattered reports ofweevil problems in south India (Viswanath 1976),Thailand (Nanthachai 1985) and Indonesia (Kusomo &Sunaryono 1985) are not supported by population oryield loss data and lack confirmation. Nevertheless,reviews by Viswanath (1976), Geddes & Iles (1991),Waterhouse (1993), and Gold (1998b) suggest thatbanana weevil is an important pest in Malaysiaand of moderate importance in parts of Indonesia,the Philippines, Sri Lanka, and Vietnam. Hasyim (pers.comm.) also believes the weevil to achieve pest statusin parts of Indonesia. An understanding of weevil peststatus and population dynamics in Asia will be criticalfor the development of a classical biological controlprogramme.

5. Yield losses to banana weevil

Yield loss to banana weevil may be affected by anumber of inter-related factors including weevil den-sity, clone, age of stand, environmental conditions(elevation, rainfall, soil type) management practices,and presence of other stresses. In Uganda, bananaweevil damage was greater: (1) on plantains (AAB)and highland banana (AAA-EA) than on Pisang awak(ABB) or Ney Poovan (AB); (2) on farms not employ-ing sanitation practices; and (3) at elevations under1400 masl (Gold et al. 1994a, 1997). Farmers arewell aware that weevil damage is more important inolder stands and on-station trials have shown increas-ing yield losses over time (Rukazambuga et al. 1998,2002). Thus, single cycle yield loss trials may bemisleading.

Yield losses attributed to banana weevil rangefrom 0% to 100% (Table 4). Methods for estimat-ing yield loss include field observations on the pro-portion of dead suckers, plant loss through mortality,toppling and snapping (Liceras et al. 1973; Ambrose1984; Sengooba 1986; Ingles & Rodriguez 1989;Sikora et al. 1989; Marcelino & Quintero 1991)and controlled trials quantifying bunch weight oryield (tonnage/ha/year) reductions (Reinecke 1976;Sponagel et al. 1995; Rukazambuga 1996; Mestre &Rhino 1997; Rukazambuga et al. 1998, 2002).Unfortunately, it is often unclear how yield loss fig-ures presented in the literature were derived and whatthey might represent (i.e. a single trial or losses withina region). In some cases, estimates represent combinedloss to weevil and nematodes (Roman et al. 1983;Sengooba 1986; Sikora et al. 1989; Gold et al. 1998b),while in some reports it is difficult to tell if yield lossesincluded nematode effects or not (Montellano 1954;Ingles & Rodriguez 1979; Job et al. 1986; Marcelino &Quintero 1991).

In yield loss trials on highland banana in Uganda,Rukazambuga et al. (1998, 2002) related levels ofweevil damage in the central cylinder to plant growth,maturation rates, and yield. Weevils were releasedat the base of banana mats 9 months after planting.Weevil populations, corm damage, plant growth, andyield were assessed over four crop cycles. Plant losswas attributable to weevils if dead, toppled or snappedplants showed heavy signs of weevil attack. In one trial,damage to the central cylinder increased from 4% inthe plant crop to 17% in the third ratoon (Table 5a)(Rukazambuga unpubl. data). Root necrosis was low


Table 4. Reported levels of yield loss to banana weevil C. sordidus

Country Yield loss (%) Clone Methods Reference

Latin AmericaBrazil 30 Moreira

20–50 Gallo (1978)Bahia Abandoned Observ. Arleu & Neto (1984)Colombia? Up to 80∗ Marcelino & Quintero (1991)

Up to 60 Castillon (2000)Cuba 22–34 Trials Reinecke (1976)

>19 Trials Calderon et al. (1991)20 Trials Masso & Neyra (1997)

Ecuador 20–40 Champion (1975)Honduras 8–26 Roberts (1955)Trials 0 Trials Sponagel et al. (1995)Peru 48∗ Liceras et al. (1973)Puerto Rico 30–70∗ Ingles & Rodriguez (1989)

902 Trials Roman et al. (1983)

AfricaCameroon 20–90 Lescot (1988)Congo Up to 90 Ghesquiere (1925)Ghana 25–90 AAB Gorenz (1963)Ghana 33 AAB Udzu (1997)Kenya 24–90 AAA-EA Trials ICIPE (1991)

16–53 AAA-EA Trials Ngode (1998)Up to 100 AAA-EA Koppenhofer (1993a–c)22–76 AAA-EA Trials Musabyiamana (1999)

TanzaniaKagera 30 AAA-EA Walker et al. (1983)

30∗,2 AAA-EA Sikora et al. (1983)15 Trials Uronu (1992)

Uganda 5–44 AAA-EA Trials Rukazambuga et al. (1998)40–502 AAA-EA Trials Gold et al. (1998)

Central 20–60 AAA-EA Damage1 Gold et al. (1999)Masaka Up to 100∗ AAA-AE Observ. Sengooba (1986)Rakai Up to 100∗ AAA-AE Observ. Sengooba (1986)

>50% AAA-AE Observ. Sebasigari & Stover (1988)West Africa 5–44 AAB Trials Sery (1988)

Asia/PacificIndia 352 Trials Job et al. (1986)Tonga <10 Damage2 Crooker (1979)

30–60 Damage2 Englberger & Toupu (1983)Up to 80∗ Observ. Pone (1994)

∗Percentage plant loss due to toppling and snapping attributable to weevils.1Estimated from Rukazambuga et al. (1998) data sets.2Composite weevil and nematode.

in all cycles suggesting that effects were most likelydue to the weevils alone.

The effect of damage was greater on bunch weightthan on plant growth and rate of development.Moderate to heavy banana weevil attack reduced thenumber of functional leaves at flowering, plant girth,plant height, and maturation period (sucker emergenceto harvest). Plant loss attributable to banana weevilattack in two trials increased from 4% in the plant

crop to 29% in the third ratoon (Table 5b). The cumu-lative effect of heavy damage sustained over severalcrop cycles resulted in greater reduction in bunchweight than that inflicted by similar levels of damagein a single cycle (i.e. by weakening the mat’s cormleading to weaker followers). Weevil damage >10%reduced bunch weights on harvested plants by 20–45%(Table 5c), while yield losses increased from 9% in thefirst ratoon to 48% in the fourth cycle.


Table 5. Effect of banana weevil damage in the central cylinder onplant loss, bunch weight and related yield loss in a trial (cv Atwalira)at the Kawanda Agriculture Research Institute, Kampala, Uganda1991–1995

Damageclass

Plantcrop

Ratoon cycle

First Second Third

a. Number of plants displaying level of damage0–5% 74% 38% 24% 6%

>5–10 19 31 30 14>10–15 5 18 19 18>15–20 2 4 8 10>20 — 8 19 52Mean damage 4.2% 8.2% 10.8% 16.9%

b. Plants lost without producing harvestable bunches0–5% 2 5 10 0

>5–10 4 13 14 0>10–15 2 5 14 3>15–20 4 3 6 7>20 — 13 43 80Percent 3 9 19 29

c. Bunch weights by damage class (mean damage per matfor given and preceding cycles) (kg)0–5% 12.2 12.9 16.8

>5–10 10.1 11.7 14.3>10–15 9.8 9.3 11.2>15–20 7.1 9.9 8.3>20 7.0 7.4 9.1

d. Yield loss by cycle

PC Crop cycle

1 2 3

Expected yield1 4153 5175 5130 5127Actual yield 3961 4709 4281 2896Yield loss 5% 9% 17% 44%

PC: Plant crop.Adopted from Rukazambuga et al. (1998).

In a second trial employing four treatments to createdifferent levels of host plant vitality, overall yield lossin the trial increased from 6% in the plant crop to 21%in the third ratoon (Rukazambuga et al. 2002). In thefourth cycle, yield losses were 27% each in the mosthighly stressed (i.e. intercropped with finger-millet)and most vigorous-growing banana (i.e. mulched withgrass). This translated into a loss of 2.5 tonnes/ha inthe intercrop and of 6.3 tonnes/ha in the mulch. Thesedata suggest that banana weevil can be an importantconstraint in well-managed banana stands.

In Honduras, Sponagel et al. (1995) attempted toassess weevil pest status through the use of pesticidechecks. Chemical applications resulted in 48–80% pop-ulation decreases compared to the control. However,the reductions in weevil populations was not reflectedin either reductions in damage or increases in yield.Sponagel et al. (1995) concluded that the weevil was

not an economic pest in Honduras. However, it isunclear how this conclusion was reached since cormdamage was similar among treatments. The data dosuggest the possibility of lag effects between killingof adults in established plantations and reducing cormdamage.

Part 2: Integrated Pest Management ofBanana Weevil

Current research results suggest that no single con-trol strategy will provide complete control for bananaweevil. Therefore, a broad IPM approach, combin-ing a range of methods, might offer the best chancefor success in controlling this pest. The componentsof such a programme include habitat management(cultural control), biological control, host plant resis-tance, botanicals, and (in some cases) chemical control.

XI. Habitat Management (Cultural Control)

Habitat management offers a first line of defenceagainst herbivores (Altieri & Letourneau 1982) bycreating an environment which reduces pest immigra-tion and/or encourages reduced tenure time and emi-gration, promotes host plant vigour and tolerance ofpest attack, and/or is unfavourable to pest build-up.For banana weevil control, habitat management optionsinclude the use of clean planting material, selectionof cropping systems, improved agronomic practices topromote plant vigour, management of crop residues(i.e. sanitation), and trapping (Table 6). Peasley &Treverrow (1986) have summarised this approach as‘start clean, stay clean’.

1. Clean planting material

The use of clean planting material can reduce initialbanana weevil infestations and retard pest build-up forseveral crop cycles. At the same time, it can protectnew banana stands against nematodes and some dis-eases. Infested planting material provides the princi-pal entry point of banana weevils into newly plantedfields. Suckers used as planting propagules often con-tain weevil eggs, larvae, and, occasionally, adults.Removing these weevils from planting material elim-inates the most important source of infestation in newplantations. The insect’s low fecundity and slow popu-lation growth further suggest that a reduction in initialinfestation level might result in lower damage to newly


Table 6. Cultural practices for control of banana weevil: Literature recommending, reviewing or reporting research with favourable results

Crop rotation or fallow to clean fieldsFroggatt (1924), Pinto (1928), Aguero (1976), Greathead et al. (1986), Jones (1986), Pinese (1989), Sikora et al. (1989), Seshu Reddyet al. (1993, 1998), Pinese & Piper (1994), Pone (1994), Price (1994), Stanton (1994), Nkakwa (1999)

Clean planting material1. Tissue culture plantlets

Peasley & Treverrow (1986), Pone (1994), Aranzazu et al. (2001)2. Selection of clean suckers

Froggatt (1925), Pinto (1928), Sein (1934), Hargreaves (1940), Weddell (1945), Haarer (1964), Saraiva (1964), Nonveiller (1965),Gordon & Ordish (1966), Trejo (1969), Wardlaw (1972), de Villiers (1973), McNutt (1974), Aguero (1976), Franzmann (1976),Suplicy Filho & Sampaio (1982), Jones (1986), Peasley & Treverrow (1986), Williams et al. (1986), INIBAP (1988b), Allen(1989), Anonymous (1989), Pinese (1989), Londono et al. (1991), Pinese & Piper (1994), Sponagel et al. (1995), Aranzazu et al.(2000, 2001), Tushemereirwe et al. (2000)

3. ParingFroggatt (1925), Veitch (1929), Sein (1934), Weddell (1945), Harris (1947), Nonveiller (1965), Gordon & Ordish (1966), Firman(1970), Wardlaw (1972), de Villiers (1973), McNutt (1974), Aguero (1976), Franzmann (1976), Mau (1981), Dawl (1985), Jones(1986), Peasley & Treverrow (1986), INIBAP (1988b), Rodriguez (1989), Londono et al. (1991), Pinese & Piper (1994), Simon(1994), Sponagel et al. (1995), Gold (1998b), Gold et al. (1998a,b), Seshu Reddy et al. (1998), Mbwana & Rukazambuga (1999),Nkakwa (1999), Aranzazu et al. (2000, 2001), Tushemereirwe et al. (2000)

4. Treatment with CreolinaAranzazu et al. (2000, 2001)

5. Water submersionGhesquierre (1924, 1925)

6. Hot water treatmentGhesquiere (1924), Sein (1934), Hildreth (1962), Barriga & Montoya (1972), Castano (1973), PANS (1973), Jurado (1974),McNutt (1974), Arroyave (1985), Stover & Simmonds (1987), Sebasigari & Stover (1988), Londono et al. (1991), Seshu Reddyet al. (1993, 1998), Pone (1994), Prasad & Seshu Reddy (1994), Simon (1994), Gold (1998b), Gold et al. (1998a,b), Mbwana &Rukazambuga (1999), Tushemereirwe et al. (2000)

7. Heat sterilisationStein (1934)

8. Quick planting to prevent re-infestationVeitch (1929), Sein (1934), Saraiva (1964), Franzmann (1976), Sponagel et al. (1995), Aranzazu et al. (2000, 2001)

Crop management1. Intercropping with coffee

Kehe (1985, 1988)2. Weeding & trash removal

Wallace (1938), Haarer (1964), Saraiva (1964), Simmonds (1966), Trejo (1969), Firman (1970), de Villiers (1973), PANS (1973),Ostmark (1974), Franzmann (1976), Jaramillo (1979), Treverrow (1985), Greathead et al. (1986), Kelly (1986), Castrillon (1989,1991), Mau (1991), Pinese & Piper (1994), Simon (1994), Vittayaruk et al. (1994), Smith (1995), Sponagel et al. (1995), SeshuReddy et al. (1998), Nkakwa (1999), Tushemereirwe et al. (2000)

3. DeleafingWallace (1938), Vilardebo (1960), Saraiva (1964), Jones (1986), Castrillon (1989, 1991), Mau (1991), Mbwana & Rukazambuga(1999), Tushemereirwe et al. (2000)

4. DesuckeringWallace (1938), Nonveiller (1965), Treverrow (1985), Tushemereirwe et al. (2000)

5. Mulch locationVarela (1993), Gold (1998b), Ssennyonga et al. (1999)

6. Deep planting of suckersKehe (1985), Seshu Reddy et al. (1993, 1999)

7. Earthing up rhizomesSwaine (1952), Saraiva (1964), Nonveiller (1965), Kehe (1988)

8. RoguingFroggatt (1925), Veitch (1929), Saraiva (1964), Nonveiller (1965), Castrillon (1991)

9. Nutrient amendments and promoting plant vigourJones (1986), Sponagel et al. (1995), Bosch et al. (1996), Tushemereirwe et al. (2000), Aranzazu et al. (2001)

10. Sanitation (destruction of crop residues)Froggatt (1924, 1925), Ghesquierre (1925), Pinto (1928), Hargreaves (1940), Harris (1947), Haarer (1964), Nonveiller (1965),Gordon & Ordish (1966), Simmonds (1966), Firman (1970), de Villiers (1973), PANS (1973), McNutt (1974), Nanne & Klink(1975), Aranda (1976), Vilardebo (1977), Crooker (1979), Mau (1981), Suplicy Filho & Sampaio (1982), Arroyave (1985),Dawl (1985), Treverrow (1985, 1993), Greathead et al. (1986), Jones (1986), Kelly (1986), Peasley & Treverrow (1986),


Table 6. (Continued)

Williams et al. (1986), Cardenas & Arango (1987), Waterhouse & Norris (1987), INIBAP (1988b), Allen (1989), Anonymous(1989), Pinese (1989), Castrillon (1989, 1991), Mau (1991), Treverrow et al. (1992), Treverrow & Bedding (1993), Treverrow &Maddox (1993), Varela (1993), Pinese & Piper (1994), Seshu Reddy et al. (1994, 1998), Simon (1994), Stanton (1994), Vittayaruket al. (1994), Smith (1995), Sponagel et al. (1995), Bosch et al. (1996), Mestre (1997), Dochez (1998), Gold (1998b), Mbwana &Rukazambuga (1999), Nkakwa (1999), Aranzazu et al. (2000, 2001), Tushemereirwe et al. (2000)

11. Burying infested plants and rhizomesGhesquiere (1924), Simmonds (1966)

Trapping1. Residues

Knowles & Jepson (1912), Pinto (1928), Sein (1934), Hargreaves (1940), Harris (1947), Wolcott (1948), Cuille (1950),Yaringano & van der Meer (1975), Mitchell (1980), Arleu & Neto (1984), Arleu et al. (1984), Londono et al. (1991), Anonymous(1992), Koppenhofer et al. (1994), Masanza (1995), Ndege et al. (1995), Seshu Reddy et al. (1995), LeMaire (1996), Gold (1998),Ngode (1998), Nkakwa (1999), Tushemereirwe et al. (2000)

2. Residues treated with pesticidesVeitch (1929), Whalley (1957), Bullock & Evers (1962), Sotomayor (1972), Yaringano & van der Meer (1975), Cardenas &Arango (1986), Treverrow et al. (1992), Cerda et al. (1994), Masanza (1996), Aranzazu et al. (2000, 2001)

3. Residues treated with entomopathogensMesquita (1988), Budenberg et al. (1993a), Kaaya et al. (1993), Carballo & Lopez (1994), Contreras (1996), Masanza (1996),Braimah (1997), Nankinga (1997, 1999), Nankinga & Ogenga-Latigo (1996), Nankinga et al. (1999), Aranzazu et al. (2000, 2001)

4. Residues treated with entomopathogenic nematodesSchmitt et al. (1992), Treverrow (1994), Aranzazu et al. (2000, 2001)

5. Enhanced trapping with semiochemicalsBudenberg et al. (1993a), Cerda et al. (1994), Ndiege et al. (1996a,b), Jayaraman et al. (1997), Alpizar et al. (1999), Tinzaara et al.(1999b), Tushemereirwe et al. (2000)

planted fields over extended periods of time. As a result,the use of clean planting has been widely recognisedand promoted.

Re-infestation remains a critical concern. The periodof protection afforded by the use of clean plantingmaterial will vary by cultivation practice. The mostpronounced effect will occur when clean material isplanted in isolated sites with no recent history ofbanana production. For example, Froggatt (1925) advo-cated the use of clean planting material and recom-mended against planting near infested stands or in fieldswhere holdover populations of weevils remained fromprior plantings. Previously infested fields can be rid ofweevils by crop rotation or fallowing (Table 6). If pos-sible, the old corms should be removed (Seshu Reddyet al. 1993; Stanton 1994). Data on adult survival inthe absence of food sources suggest this period shouldbe at least 4–6 months (Froggatt 1924; Peasley &Treverrow 1986; Pinese 1989; Seshu Reedy et al. 1998)and possibly up to 24 months (Treverrow 1985).

In some regions, clean, isolated fields may not beavailable and the ability to take land out of banana pro-duction (i.e. crop rotation or fallowing) may be limited.In Uganda, where highland cooking banana (AAA-EA)is the preferred staple and an important component offood security, high population density and land pres-sure preclude the use of isolated fields (Gold et al.1999b). As a result, gap filling in infested fields and

planting of new stands proximal to established, infestedbanana stands is common. Under such conditions, theimpact of clean planting material will be reduced (Sein1934; McIntyre et al. 2002).

A number of methods have been proposed for clean-ing planting material of weevils. These include theuse of tissue culture plantlets; selection of weevil-freesuckers; paring, immersion in cold water, hot watertreatment, and/or heat sterilisation of suckers; and theuse of entomopathogens.

The use of tissue culture plantlets as a means ofbanana weevil control has been recommended byPeasley & Treverrow (1986) and Pone (1994). Unlikeother methods, tissue culture plants are likely tobe 100% free of banana weevils and nematodes at thetime of planting. Once established in the field, it isunclear whether tissue culture plants are more or lesssusceptible to banana weevils than plants grown fromsuckers.

Tissue culture material is widely used for com-mercial banana production in Latin America, Asia,and Africa for pest and disease control. For exam-ple, Dochez (1998) found that 77% of surveyed com-mercial farmers on the south coast of Kwazulu/Natal,South Africa used tissue culture plantlets. AlthoughSeshu Reddy et al. (1998) suggest that tissue cultureis beyond the reach of most small-scale farmers insub-Saharan Africa, production and dissemination of


tissue culture plantlets is currently being promoted inBurundi, Kenya, Rwanda, and Tanzania.

Selection of weevil-free planting material by care-ful observation of plants in the field has been widelyrecommended (Table 6). This entails selecting suckersfrom fields with no or low weevil infestation and/orrejecting suckers that appear to have weevil damage.However, banana suckers may carry eggs and/or early-instar larvae, which are not easily detected by visualobservation. In areas where weevil problems are severe,most farms may be infested and farmers may not be ableto choose clean suckers.

Paring, or removal of the outer surface of the corm,has also been widely recommended (Table 6). Paringcan expose weevil galleries and allow the farmer toreject heavily damaged suckers. Removal of all leafsheaths and paring of the entire corm will elimi-nate most weevil eggs and many first-instar larvae.However, later instar larvae are often deep within thecorm and will not be removed by paring (Hildreth 1962;Reinecke 1976; Arroyave 1985).

Paring the entire corm normally entails removal ofall the roots. This method has also been recommendedas a means of nematode control (Speijer et al. 1995).Concerns about viability of pared suckers have beenraised by Hord & Flippin (1956) and Coates (1971).Although pared suckers may suffer under conditions oflow soil moisture, in Uganda growth of pared suckersin field trials is usually satisfactory (Gold et al. unpubl.data).

Ghesquiere (1924, 1925) suggested that submerg-ing suckers in water for 2 days would kill all weevileggs, larvae, and adults. In contrast, Froggatt (1924),Gettman et al. (1992), and Minost (1992) reported thismethod as ineffective. Simmonds (1966) suggested thatsoaking required 3 weeks to eliminate weevils, whichwas both impractical and would cause loss of plant-ing material. As a result, this method is rarely recom-mended.

Hot water treatment to kill weevil eggs and larvaewas first recommended in the 1920s and continues to bepromoted (Table 6). Ordinarily, the corms are pared andthen completely submerged in hot water. Sein (1934)reported that placing suckers in boiling water for 1 minkilled all weevil eggs and surface larvae, while heatsterilisation at 43◦C for 8 h eliminated larvae deeperwithin the corm. Arroyave (1985) summarised recom-mendations for hot water treatments in Latin America(i.e. Hildreth 1962; Barriga & Montoya 1972; Jurado1974; Castano 1983) in which prescribed temperaturesranged from 54◦C to 60◦C and submersion times from 8

to 20 min. No data on treatment efficacy were presentedfor any of these studies.

The use of some hot water treatment regimes (52◦Cfor 27 min or 54–55◦C for 20 min) is also a highly effec-tive control against banana nematodes (Seshu Reddyet al. 1993; Prasad & Seshu Reddy 1994; Speijeret al. 1995). These temperatures have been suggestedfor concurrent management of weevils and nematodes(Seshu Reddy et al. 1993; Prasad & Seshu Reddy1994).

In Hawaii, Gettman et al. (1992) reported greaterthan 99% mortality of weevil eggs and larvae whensuckers of dessert bananas (AAA) were placed in awater bath of 43◦C for 3 h. Arroyave (1985) testedhot water treatments (52–55◦C for 15 min) for clean-ing plantain suckers in Colombia and found that lar-vae within the interior of the corm survived. Sheattributed this to failure of the heat to penetrate throughthe corm and suggested that larger suckers would havegreater larval survival. These larvae would then providefocal points for re-infestation. Similarly, Peasley &Treverrow (1986) and Treverrow et al. (1992) reportthat hot water baths are not effective at killing larvaedeep within the corm.

Efficacy of paring and hot water treatment in killingweevil eggs and larvae was also tested in Uganda (Goldet al. 1998a). Paring removed >90% of weevil eggs buthad little effect on weevil larvae. Mortality of bananaweevil immatures was also recorded after immersionof infested banana suckers in four hot water regimes:43◦C for 2 h, 43◦C for 3 h, 54◦C for 20 min, and 60◦Cfor 15 min. All hot water treatments resulted in 100%mortality of eggs. However, only hot water baths of43◦C for 3 h resulted in high mortality (i.e. 94%) ofweevil larvae. Larval mortality in other treatments was26–32%. Larval survival was considerably higher inthe central cylinder than in the cortex.

Banana weevils are attracted to cut corms and mayquickly re-infest pared or hot water treated plantingmaterial if these are left in an area exposed to weevils.Therefore, quick planting of treated material is alsorecommended (Table 6). Planting material may alsobe protected by pesticide (Rukazambuga 1996), neem(Musabyiamana 1999) or entomopathogen (Godonou1999) dips or applications in the planting hole. Inrecent years, research has been conducted on micro-bial control of banana weevil to reduce re-infestationrates and prolong the benefits of clean planting mate-rial. Griesbach (1999) initiated research on the use ofendophytes that may be inoculated into tissue cultureplantlets and reduce weevil infestation (see below).


In Ghana, Godonou (pers. comm.) failed to establishisolates of B. bassiana as endophytes, but successfullydemonstrated that application of B. bassiana formula-tions into planting holes could reduce weevil attack ofsuckers during the crop establishment phase (Godonou1999; Godonou et al. 2000).

Field trials in Uganda compared weevil andnematode populations, plant growth and yield in(1) untreated suckers (controls); (2) pared corms;(3) pared and hot water (54◦C for 20 min) treated corms(Gold et al. 1998b). Weevil numbers were lower intreated material than in control plots for 11–27 months.Weevil damage levels in controls were 1.7–3 timeshigher than in plots grown from treated planting mate-rial for the plant crop. However, all treatments dis-played similar levels of weevil damage in the firstratoon. Hot water treatment had little advantage overparing for controlling weevil but afforded excellentnematode control for the duration of the trial. Plantsgrown from treated material had faster maturationrates and lower levels of plant loss (2–4%) due to peststhan untreated plants (21–34%). Thus, yield per ha was1.4–2.8 times higher in plots grown from treated thanfrom untreated material for the first 28 months of thetrials, even though there were no treatment effects onbunch size.

Farmer adoption of clean planting material tech-nologies clearly varies from region to region and evenamong sites within regions. Where tissue culture isnot available or affordable, selection of clean suckersshould be straightforward. However, many farmers willreject only the most seriously damaged suckers. Paringto remove weevil eggs and expose larval damage hasnot been widely adopted by farmers in East Africa.Many farmers believe that suckers will not performwell following removal of most or all of the root system.In Tanzania, for example, Taylor (1991) reported thatfarmers viewed the recommendation of corm paringwith ‘extreme disbelief’.

Implementation of hot water baths for control ofbanana weevils and nematodes requires investment in ahot water tank and a heating source (e.g. electricity, gasburner, wood). As a result, adoption by resource-poorfarmers may be limited (c.f. Ssennyonga et al. 1999;C. Kajumba unpubl. data). Moreover, it is unlikely thatfarmers would adopt hot water baths of 3 h, as requiredfor highly effective weevil control. Additionally, con-trol of the proper temperature is important becauseof the delicate balance between killing pests anddamaging the plant (Castano 1983; Seshu Reddy et al.1998). Therefore, Gold et al. (1998a) suggested that

paring alone might be adequate for reducing weevils,although hot water baths would continue to be preferredfor nematode control.

In a survey of farmers in Kisekka subcounty inMasaka district, Uganda, 77% of farmers tried to selectuninfested suckers, 37% cut out damaged sections ofthe corm, while 46% rogued severely infested plants(Ssennyonga et al. 1999). Few farmers (3%) wereaware of paring and less than 2% regularly did this.Twelve percent tried to obtain material for cultivarsthat they deemed ‘tolerant’ of weevil attack.

In summary, infested planting material provides theprincipal entry point of banana weevils and nematodesinto newly planted fields. Use of clean planting mate-rial reduces initial weevil numbers and, thereby, retardspopulation build-up. Tissue culture plants offer onemeans of assuring pest-free planting material althoughproduction capacity, costs and means of disseminationare limiting factors in some countries. Alternatively,paring the corm removes most eggs and exposesdamage of heavily infested plants that may then berejected. Hot water treatment (20 min at 55◦C) fur-ther reduces weevil numbers by killing larvae withinthe corm. However, neither paring nor hot water treat-ment completely eliminates banana weevil. The factthat Gold et al. (1998b) no differences in infesta-tion levels between bananas grown from treated anduntreated planting propagules in first ratoon cropsputs into question the use of clean-planting materialin subsistence systems where stands are expected tolast many years. Thus, the use of clean material pro-vides initial protection to a banana stand, but ulti-mately needs to be integrated with other weevil controlmethods.

2. Cropping systems and crop management

The employment of multiple cropping systems as ameans of controlling banana weevil may be limited.Mixed cropping systems often result in lower insectpressure by reducing immigration rates, interferingwith host plant location and increasing emigrationrates (Altieri & Letourneau 1982; Risch et al. 1983).However, banana weevils are sedentary insects that livein perennial systems in the presence of an abundantsupply of hosts. Moznette (1920) concluded that theweevils only move out of overcrowded or depletedresources, a condition unlikely to occur in bananaplantations.

Cropping systems effects on banana weevil levelswould most likely be through changes in host plant


quality and/or microclimates (i.e. factors influencingsoil moisture). To date, there is little evidence of effec-tive natural enemies whose action might be enhancedby crop diversification in banana stands (c.f. Root 1973)with the possible exception of myrmicine ants (seebelow). At the same time, the banana weevil’s limitedmobility suggests that intercropping will have minimaleffect on banana weevil immigration and emigrationrates or tenure time.

Selection of cropping systems that may discouragebanana weevils include intercropping with insect repel-lent crops or green manures. Kehe (1985, 1988) sur-veyed farms in Cote d’Ivoire and found that plantainsmixed with older coffee stands (i.e. >5 years) had lowincidence of weevil attack (mean CI = 6%), whileplantain mixed with younger coffee plants (CI = 91%),with cacao (CI = 88%) or with annual crops (CI =79%) all suffered high levels of attack. He postulatedolder that coffee plants produced sufficient caffeine toserve as an effective insecticide or feeding inhibitor.Kehe (1988) suggested that the caffeine is released intothe soil and is (presumably) absorbed into the plantwhere it is effective against weevil larvae. Although,Sarah (1990) found that spreading coffee mulch at thebase of banana mats had disappointing results, manyfarmers in Masaka district, Uganda believe applicationof coffee husks does reduce weevil levels (Ssennyongaet al. 1999). Quantitative studies will be needed toverify this hypothesis.

In Tanzania, a series of trials on intercropping andbanana weevils failed to produce viable crop mix-tures that would both reduce weevil damage and pro-duce satisfactory banana yields (Uronu 1992). Reducedweevil populations occurred only in mixtures withsweet potato where the bananas were badly stuntedby intercrop competition. In Uganda, intercrops ofgreen manures with reported insecticidal properties(i.e. Canavalia, Mucuna, Tephrosia) had no effect oneither weevil adult numbers or on damage (McIntyreet al. 2002). Application of Crotalaria as mulch hadno influence on either weevil numbers or damage(McIntyre et al. unpubl. data). Similarly, Salazar (1999)found no significant effect of Mucuna on banana weevilpopulations in Puerto Rico.

It has often been suggested that banana weevil is agreater problem in poorly managed stands (see above).For example, farmers in central Uganda attributedincreasing weevil problems to reduced labour avail-ability and concomitant reductions in managementattention (Gold et al. 1999a,b). Conversely, higher lev-els of management might serve to reduce weevil pest

status. For example, Jones (1986) and Sponagel et al.(1995) suggest that practices encouraging vigorousbanana growth might allow the plant to arrest or tolerateweevil attack.

Weeding, removal of trash from the base of the mat,deleafing and desuckering have all been reported asmeans of eliminating shelters and hiding places forweevils or making the environment at the base of themat less favourable to ovipositing females (Table 6). InKisekka subcounty, Masaka district, Uganda, nearly allfarmers deleafed to reduce wind damage. Two-thirds ofthese farmers also reported that they also removed oldleaves to help control weevils; most felt that this methodwas moderately to very effective in reducing weevildamage (Ssennyonga et al. 1999). However, no dataare available in Masaka district or elsewhere to showthe relationship between these forms of crop sanitationand weevil damage levels.

Recent work has demonstrated that grass mulchesmay increase weevil damage by creating a morefavourable environment (i.e. cool. moist conditions)for adult weevils (Price 1994; Rukazambuga 1996;Braimah 1997). In Tanzania and Uganda, some farm-ers mulch away from the base of the mat as ameans of reducing weevil infestations (Varela 1993;Ssennyonga et al. 1999; Gold et al. 1999d). For exam-ple, in Kisekka subcounty, 35% of farmers believedthat mulching away from the base of the mat helpedcontrol weevils, while 19% practised this method.In on-station and on-farm trials, weevil populationsand damage were consistently higher in mulched thanin unmulched plots, while mulch location (i.e. toor away from the base of the mat) had little effecton either weevil numbers or damage (Gold et al.unpubl. data).

Deep planting and earthing up (Table 6) havebeen recommended to render the corm inaccessi-ble to ovipositing females and to prevent high mat.Seshu Reddy et al. (1993) planted cooking bananasat depths of 15, 30, 45, and 60 cm in drums andreported that shallow planted suckers were more proneto attack, although some weevils were able to findthe deepest planted suckers. The longer-term effectsof deep planting and earthing up are unclear. Abera(1997) showed that weevils freely oviposit in leafsheaths, while Masanza (unpubl. data) found increasedoviposition on buried versus unburied corms dur-ing dry seasons. In addition, deep planting is likelyto affect oviposition levels in the plant crop onlyas the corm will move towards the soil surface inratoon crops.


Roguing of obvious weevil-attacked plants has alsobeen recommended (Table 6). However, only the mostseverely attacked plants can be identified by obviousexternal symptoms (c.f. Rukazambuga 1996) and it ismost likely that roguing would have a limited effect onweevil population and damage levels in the remainderof the banana stand.

3. Crop sanitation

Following harvest, crop residues may serve as shel-ters for adults (Gold et al. 1999d) and oviposition sitesfor females (Abera 1997). For example, Gold et al.(1999d) found 25–32% of adult weevils associatedwith prostrate (i.e. cut or fallen) residues, while another10–12% were found in standing stumps.

For some clones, banana weevil damage is muchhigher on residues than on growing banana plants. InEcuador, Vilardebo (1960) reported that 75–80% ofweevil attack in Gros Michel was directed towardsresidues, while most attack in Petite Naine or Robustaclones was against the growing plant. Under condi-tions of low weevil pressure in Australian Cavendishplantations, weevil activity was almost exclusively incorms of harvested plants (Treverrow et al. 1991).Treverrow & Bedding (1993) also found that 60% ofthe weevils emerged from residues. Ostmark (pers.comm.) felt that most attack against Cavendish inCentral America came after harvest; therefore, plan-tations were able to tolerate very high levels of weevilswithout suffering yield loss. However, Stanton (1994)has suggested that heavy attack of old corm can weakenanchorage and lead to toppling.

The attraction of adult weevils to cut corms makesresidues especially attractive. Rukazambuga (pers.comm.) found >200 eggs on a single stump of thesusceptible highland cooking cultivar Atwalira. InIndonesia, extensive oviposition, reflected in largenumbers of early-instar larvae, was observed in cormdisk traps placed directly on the soil (C. Gold pers.observ.). Moreover, recently harvested Pisang awakand stumps to be largely weevil-free, while up to100 larvae per residue were observed on harvested andcut plants left prostrate on the soil (C. Gold pers. observ.1997). In Uganda, Gold & Bagabe (1997) reportednegligible damage on recently harvested Kisubi (AB)in Uganda, but extensive tunneling in its residues.Heavy attack on residues might also reflect increasedsurvivourship as compounds conferring resistance insome clones might break down after harvest. For exam-ple, Kiggundu found that weevils freely oviposited

on Kisubi yet damage, prior to harvest, was light.Preferential attack and/or increased larval success onprostrate residues may result from the exposure of thetrue stem to ovipositing weevils, whereas in standingplants and stumps, the first-instar larvae must oftentunnel through the pseudostem before reaching itspreferred food source.

It is widely believed that destruction of crop residues(splitting of harvested pseudostems and/or removalof corms) eliminates adult refuges, food sources, andbreeding sites, lowers overall weevil populations andreduces damage on standing plants in susceptibleclones (Table 6). While destruction of residues willkill any eggs and larvae in them, it is also possi-ble, that the residues may serve as traps that drawgravid females away from growing bananas (Peasley &Treverrow 1986; Waterhouse & Norris 1987; Gold1998a). Treverrow (1985) and Allen (1989) recom-mend placement of 60-cm lengths of cut pseudostemsin banana stands to lure ovipositing females away frombanana mats. This material quickly dries out, lead-ing to wastage of any eggs placed in these residues.Ghesquiere (1924, 1925) suggested destroying or evenburying old stems and corms. Hargreaves (1940) advo-cated cutting residues at ground level, splitting pseu-dostems and spreading them as a mulch, coveringcorms with compact soil or, if heavily infested, remov-ing and chopping them. These recommendations arenow widespread (Table 6).

Nevertheless, some farmers and researchers believethat nutrients and water move from residues to fol-lowers and that up to 1.5 m of pseudostem should beleft ‘in situ’ (Peasley & Treverrow 1986; Treverrowet al. 1992; Smith 1995; Sponagel et al. 1995). INIBAP(1988b) recommended cutting residues at ground levelin East Africa (to prevent weevil larvae moving fromthe mother plant to the follower), while leaving residuesat 1 m in West Africa.

The value of sanitation as a means of weevil controlhas been disputed. For example, Peasley & Treverrow(1986) and Treverrow et al. (1992) suggest that crophygiene (i.e. sanitation) is the long-term key to weevilcontrol and that without it all other control measuresare pointless. Nanne & Klink (1975) report that sani-tation can drastically reduce weevil populations. Jones(1986) indicated that control is mostly linked to san-itation. Gold et al. (1997) determine weevil levels on50 farms in Ntungamo district, Uganda and found thatsanitation had more impact on weevil pest status thanany other agronomic practice, while in Masaka districtTinzaara et al. (unpubl. data) found lowest levels of


weevils on farms employing the highest levels of sanita-tion, In contrast, Jones (1968) suggested that sanitationrequires too much labour, while Ostmark (pers. comm.)felt that weevils were not serious pests in commercialplantations to and, therefore, sanitation was worthlessand not worth the effort. Much of this debate is spec-ulative, based on causal observations, perceptions ofweevil pest status, and beliefs on weevil populationdynamics. Unfortunately, there have been virtually nodata from controlled studies on the role of crop sanita-tion in weevil population dynamics and related damage.

During the rainy season, Masanza (1999) found thatcorms cut 5 cm above the soil surface had twice as manyeggs as when cut at the soil surface and four timesas many eggs as corms cut 5 cm below the soil sur-face and covered with soil. In contrast, during the dryseason buried corms had three times as many eggs ascorms cut at or above the soil surface. Masanza (1999)attributed these seasonal differences to shifts in soilmoisture profiles.

Masanza (1999) also looked at attraction and accep-tance of different types of highland cooking bananaresidues in the laboratory. Consistent with the findingsof other researchers, weevils were more attracted toand oviposited more on corm material than on pseu-dostems. Surprisingly, the weevils were more attractedto corms >30 days after harvest than freshly cut corms.Under field conditions, banana weevils oviposited oncorms up to 120 days after harvest, although femalesplaced four times as many eggs on fresh corms as those>30 days old. Eclosion rates were independent of cormage, although larval/pupal survivourship was greaterand the combined stage duration shorter on fresh corms(11.5% survivourship; median 38 day duration) than on14–30 days old corms (7.5%; 43 days) or >30 days oldcorms (4.5%; 47 days). Pupal weights were also greaterfor larvae reared on fresh corms. Throughout Uganda,many farmers remove corms many weeks, rather thanimmediately, after harvest. Masanza’s (1999) data sug-gest that removal of fresh corms may be far moreeffective in reducing weevil numbers.

In a survey of commercial Cavendish plantations inthe south coast of Kwazulu/Natal, South Africa, allfarmers believed residues served as breeding groundsfor weevils (Dochez 1998). However, only 53% of thefarmers were willing to remove them due to labourcosts and the belief that the residues were beneficial tothe growth of followers.

During rapid rural appraisals at 25 sites inUganda, many farmers recognised the theoreticalvalue of crop sanitation (widely recommended by

extension services), but few practised it because it wasseen as costly and time consuming (Gold et al. 1993). Ina second study, farmers in central Uganda attributed therecent decline in highland cooking banana productiv-ity, in part, to increasing damage to banana weevil thatwas aggravated by lack of field sanitation (Gold et al.1999b). The abandonment of field sanitation practiceswas related to a relaxation of government by-laws(held over from the colonial period) and to loweravailability/increasing costs for external labour.

Farmer interviews on crop sanitation practices werealso conducted with farmers in Masaka and Ntungamodistricts, Uganda (Ssennyonga et al. 1999; Masanzaet al. unpubl. data). In both sites, farmers implementeda wide range of residue management practices rang-ing from cutting residues a few centimetres belowground level to up to 1 m above the collar. Cut residueswere mostly left intact, chopped or split and spread asmulch. Some farmers covered corms with compactedsoil to prevent weevil oviposition, while others cov-ered corms with leaves in a modified disk on stump trap.Nevertheless, the majority of farmers implemented low(i.e. sporadic) levels of sanitation while only a fewsystematically destroyed crop residues.

4. Trapping adult weevils

Trapping to monitor weevil populations and the effi-cacy of different types of traps has been discussedabove. The use of trapping as a means of control-ling banana weevils has also been recommended bymany workers (Table 6), although this approach hasbeen controversial (INIBAP 1988a,b; Gold et al. 1993).Modifications (e.g. addition of chemicals, biopesti-cides, or semiochemicals) on basic trapping meth-ods have also been proposed. Jayaraman et al. (1997)and Alpizar et al. (1999) suggested that mass trap-ping with semiochemicals could overcome the weevil’slow fecundity and slow population build-up and leadto successful control, while Braimah (1997) offeredthat the use of pseudostem traps enhanced by semio-chemicals and combined with other compatible con-trol methods (e.g. entomopathogens) holds the key tobanana weevil control. In contrast, Mestre (1997) con-cluded that the weevil is a poor candidate for mass trap-ping with semiochemicals because it is soil dwelling,sedentary, and rarely flies.

The effect of trapping on weevil populations will, inpart, reflect the intensity of trapping (trap density andtrapping frequency) and the types of materials used. Inaddition, it is likely that trapping in established fields


will result in a gradual decline in weevil numbers witha lag time required before effects are manifested inreduced damage.

Weevil reductions due to trapping have been reportedby Vilardebo (1950), Arleu & Neto (1984), Arleuet al. (1984), Koppenhofer et al. (1994), Seshu Reddyet al. (1995), Ndege et al. (1995), Masanza (1995),Ngode (1998), and Alpizar et al. (1999). However, theuse of trapping as a control of banana weevil has alsobeen considered ineffective or impractical (Roberts1955; Braithwaite 1958; Jones 1968; Ostmark 1974;Jaramillo 1979; Stover & Simmonds 1987; INIBAP1988b; Gold & Gemmell 1993; Gowen 1995).

Data from controlled field studies are largelywanting. In Honduras, Roberts et al. (1955) andOstmark (1974) report of a 2-year study in whichone million weevils were collected from 16.2 ha ofbanana (i.e. 2575 weevils/ha/month), but trap catcheswere similar at the beginning and end of the study.The authors concluded that trapping is ineffective forweevil control. However, no information was providedon trap density and weevil population levels so it is hardto determine what proportion of weevils were beingremoved.

In contrast, Yaringamo & van der Meer (1975)reported a 50% population reduction in Peru from4 months of corm trapping, but the means by whichthis reduction was determined is not clear. In Kenya,Seshu Reddy et al. (1995) also found a 50% reduc-tion in weevils captured following systematic trapping.Koppenhofer et al. (1994) ran a series of experiments tolook at the effects of pseudostem trapping (at variabletrap densities reflecting available material) on weevilnumbers. In the first experiment (weekly trapping), trapcaptures declined by 33% over an 11-week period. Ina second experiment (weekly trapping), weevil num-bers collected in traps declined by about 50% in 1 yearin one field, but showed no change in a second field.In a third trial, marked weevils were released into afield and populations were estimated using the Lincolnindex. Weevils were collected from traps on a dailybasis for 7 weeks, after which the population was esti-mated again. During this time period, weevil densityhad declined by 59–67% from the original release level.

In all of these studies, comparisons were made withinitial populations and, thus, the trials lacked propercontrols, making the results inconclusive. For exam-ple, reported weevil reductions in the Seshu Reddyet al. (1995) study and first two Koppenhofer et al.(1994) trials were interpreted from trap capture rates,which may have reflected weather conditions and trap

efficiency (Vilardebo 1973). Similarly, weevil popula-tion declines of the same magnitude as that reported inKoppenhofer et al.’s (1994) third trial have been foundfor field populations of marked and released weevils intrials where trapping was not conducted (Rukazambuga1996; Gold & Night unpubl. data).

Controlled studies to determine the efficacy ofpseudostem trapping in reducing weevil populationswere conducted under farmer conditions in Ntungamodistrict, Uganda (Gold et al. 2002b). First, a par-ticipatory rural appraisal was conducted in whichfarmers expressed concern about yield declines incooking banana that they attributed primarily tobanana weevil (Okech et al. 1996). Observationson farmers’ fields confirmed high weevil popula-tions (estimated through mark and recapture meth-ods) and damage levels on many farms (Gold et al.1997). Twenty-seven farms were then stratified on thebasis of weevil population density and divided amongthree treatments: (1) researcher-managed trapping (onetrap/mat/month): (2) farmer-managed trapping (trapintensity at discretion of farmer); and (3) controls(no trapping). In researcher-managed trapping, weevilswere collected once per month, 3 days after placementof traps.

Intensive trapping (managed by researchers) resultedin significantly lower C. sordidus damage after 1 year.Over the same period, C. sordidus numbers declinedby 61% in farms where trapping was managed byresearchers, 53% where farmers managed trapping and38% in farms without trapping; however, results var-ied greatly among farms and, overall, there was nosignificant effect of trapping on C. sordidus numbers.Moreover, there was only a weak relationship betweenthe number of C. sordidus removed and the change inpopulation density. Trapping success appeared to beaffected by management levels and immigration fromneighbouring farms.

Recent trapping studies by ICIPE in Kenya foundthat a high percentage of weevils could be removed bycontinuous intensive trapping (2 traps per mat) overseveral months (S. Lux pers. comm.). However, theamount of material required for such trapping was unre-alistic. Nevertheless, trap efficacy might be improvedby the use of semiochemicals including pheromonesand plant volatiles, by themselves or used as deliverysystems for entomopathogens.

Adoption of systematic pseudostem trappingrequires discipline and commitment on the part of thefarmer (Nonveiller 1965). Following the completionof the Ntungamo study, adoption was very low due to


the resource requirements of this method, even thoughmost farmers were convinced that trapping could bebeneficial (Gold et al. 2002b). Ndege et al. (1995)also noted that material and labour requirements mightbe beyond the means of many subsistence farmers inwestern Tanzania.

In a survey of highland cooking banana grow-ers in Kisekka subcounty, Masaka District, Uganda,Ssennyonga et al. (1999) found that 75% and 12%of farmers knew of disk on stump and pseudostemtrapping, respectively. Yet, only 15% of all farmerspractised systematic disk on stump trapping, while nofarmers implemented pseudostem trapping. In addi-tion, farmers must have realistic expectations on thebenefits of trapping. During a rapid rural appraisal inKabarole district, Uganda, farmers reported that theytried trapping for a few weeks but abandoned it whenthey failed to see immediate improvement (Gold et al.1993).

The use of enhanced trapping with semiochemicalscould result in higher rates of weevil removal at lowertrap densities and with reduced labour. The commer-cial company Chemtica International, in Costa Rica,tested lures with male aggregation pheromones andfound that a formulation, Cosmolure+, (containinga mixture of the four sordidin isomers plus plantvolatiles) was most attractive to both male and femalebanana weevils (C. Oehlschlager pers. comm.). InUganda, these lures captured as many as 18 times thenumber of weevils/day as conventional pseudostemtraps (Tinzaara et al. 1999a), while in Costa Rica,Alpizar et al. (1999) reported that pitfall traps withCosmolure+ collected 12 times as many weevils asunbaited sandwich traps. In addition, pseudostem trapsmay last for only 3–7 days and require frequent visits toremove and destroy the weevils (which enter and leavethe traps). By contrast, weevils drown in pheromonetraps, which can remain effective for up to 1 month.

Using interference studies by collecting weevilsfrom pheromone-baited pitfall traps placed at dif-ferent distances, Oehlschlager (pers. comm.) deter-mined that the optimum spacing of traps was 20 m(by contrast, Alpizar et al. (1999) estimated theeffective attractivity radius of Cosmolure+ traps at2.5–7.5 m). Based on these results, Chemtica rec-ommended a density of 4 traps/ha, placed initiallyin a single line at 20 m apart, 10 m in from border(C. Oehlschlager pers. comm.). The traps are replacedmonthly and moved 20 m further into plot giving fullcoverage after 9 months. Three assumptions are made:(1) in 1 month, the traps remove most weevils within

a 20-m radius; (2) the weevils are sedentary and rein-vasion into areas where traps have already been placedis negligible; (3) a limited number of weevils emergefrom the bananas in areas where trapping has beencompleted.

In a preliminary study to test the first of theseassumptions, Gold and Kagazi (unpubl. data) placedfive Cosmolure+ traps at the base of banana mats (trapmat) in a heavily infested stand (mat spacing at 3 m).Weevils were collected from the traps for 1 month, afterwhich the trap mat and its 8 nearest neighbours wereuprooted and remaining weevils counted. A total of395 weevils were collected from the traps, 321 weevilswere collected from the base of the five trap mats and242 weevils from 17 of 40 neighbouring mats. Extrap-olation suggests a total of 569 weevils on adjacentmats. This would indicate that the pheromone traps col-lected only 31% of the weevils within a radius of 3 m.Nevertheless, the data suggest that many weevils wereattracted by the pheromone lures, as an average of 64weevils were recovered from the mat adjacent to eachtrap compared to 14 weevils at each of the neighbouringmats.

Nevertheless, Alpizar et al. (1999) obtained verypositive results using Chemtica recommendations onCosmolure+ trap number and placement in three plan-tain fields and in one Grand Enano (AAA) stand. In theplantain systems, weevil capture rates in treated plotsremained at initial levels for 9 months and steadilydecreased thereafter, while trap captures remainedsteady in control plots. Over 18 months, damage lev-els, measured by Vilardebo’s (1973) CI, decreased from15% to 12% in the treated plots, while increasing from15% to 34% in controls. This resulted in a 25% yieldgain for the first 18 months. Similar results were foundin the Grand Enano field where treated plots had lessdamage and a 32% yield advantage.

The use of Cosmolure+ traps is now being testedin a number of countries in Latin America, as well asin Uganda, Cameroon, South Africa, India, and else-where. These studies will demonstrate the efficacy ofthe pheromone with different biotypes of the weeviland in different agro-ecological conditions. Shouldpheromones prove effective, however, the use of suchtraps will entail resolving logistics related to importa-tion, distribution, and storage, as well as monitoringthe costs and benefits to farmers.

In Kenya, ICIPE has been working on the useof kairomone traps made with processed pseu-dostem material that is then buried in the soil.High numbers of weevils are attracted to these traps


(S. Lux pers. comm.). The objective of this trap-ping system is to create ‘killing nodules’ wherebyweevils are attracted to these traps and then killedby entomopathogens (e.g. B. bassiana or Metarhiziumanisopliae) applied to the traps. If successfully devel-oped, such a system would require production anddistribution of an entomopathogen (which might bemass-produced locally) rather than a pheromone thatwould require importation from Costa Rica.

5. Adoption of cultural controls

Implementation of cultural controls of banana weevilvaries by region and may reflect a range of fac-tors including: (1) susceptibility of the predomi-nant banana clones; (2) farmer perceptions on the(potential) severity of weevil problems; (3) farmerobjectives (i.e. subsistence vs. commercial; prophylaticvs. remedial control strategies); (4) farmer resources(e.g. labour, finances, equipment), (5) farmer aware-ness of control methods; (6) extension recommenda-tions; (7) the ability of the farmer to modify methodsto suit his resources and needs; (8) farmer perceptionsof control efficacy in reducing weevil pest pressure;(9) access to inputs, e.g. clean planting material(e.g. tissue culture); (10) the length of time beforebeneficial effects might become apparent. For exam-ple, farmers in Kabarole district, Uganda abandonedpseudostem trapping because they did not see weevilreductions in a few weeks time (Gold et al. 1993),while farmers in Lwengo subcounty, Masaka districtwere disappointed with pheromone traps because ofunrealistically high expectations for immediate impact.In contrast, a majority of farmers in Kisekka subcounty,Masaka Distict, Uganda said they would be willing totest a new method for 6 months before deciding uponits value.

Some prescribed methods like deleafing or crop san-itation may be practised for agronomic, rather thanpest control purposes. For example, >75% of surveyedfarmers in the Kisekka subcounty study split pseu-dostems and/or removed stumps (Ssennyonga et al.1999) which they used as mulch. Of these, less thanhalf recognised sanitation a possible means of weevilcontrol. Other farmers removed old corms to give themat room to grow, although some felt that corms ofrecently harvested plant provided anchorage to the mat.The intensity with which weevil management practiceswere implemented varied considerably among farmersand often reflected their economic status. Farmers prac-tising systematic sanitation of banana residues tended

to sell a higher proportion of their crop and have ahigher levels of resources than those who did not.

In areas of Uganda with limited commercial oppor-tunities for banana, implementation of labour-intensivecultural controls was limited. In Ntungamo district, forexample, 75% of farmers practised little or no sani-tation and either left harvested stumps on the mat or,if cut, left them intact to rot (Masanza 1999). Twentypercent of farmers carried out sporadic sanitation inwhich they would destroy some but not all residuesor would wait until many residues were more than1 month old. Only 5% of farmers implemented sys-tematic sanitation in which most residues were fullydestroyed soon after harvest. In addition, most farmersfelt that pseudostem-intensive trapping was not a feasi-ble control strategy because of its labour and materialrequirements (Gold et al. 2001). Similarly, in centralUganda, few farmers were willing to implement labour-intensive controls against banana weevil, even thoughthey perceived this pest as a leading cause of bananadecline (Gold et al. 1999b).

XII. Biological Control with ArthropodNatural Enemies

Biological control is the combined action of a natu-ral enemy complex (parasitoids, predators, pathogens),antagonists or competitors in suppressing the popu-lation density of a pest to a level lower than wouldoccur in their absence (Debach 1964; van Driesche &Bellows 1996). Most often, biological control includesthe successful establishment of natural enemies andhas the advantages that it is ecologically sound, com-patible with most farming practices (except the useof pesticides) and requires little or no investment onthe part of the farmer. Some natural enemies mayrequire periodic augmentative releases to bolster exist-ing populations or to insure rapid dissemination intonew sites. Otherwise, successful biological controlis permanent and stabilises herbivore populations atlow levels, thereby reducing the risks of outbreaks.Even partially successful biological control would con-tribute to IPM of banana weevil since natural ene-mies are most often compatible with breeding for hostplant resistance and cultural controls (Neuenschwander1988).

Biological control efforts against banana weevilhave included the use of exotic natural enemies(classical biological control), endemic natural enemies,secondary host associations, and microbial control


(e.g. entomopathogens, endophytes, entomophagousnematodes). Microbial control agents may requirerepeated applications and may be considered as biopes-ticides, although they lack the toxic side effectsof chemical insecticides. As such, they may entailrepeated application costs on the part of the farmer.

1. Classical biological control

Classical biological control of banana weevil may bepossible. Introduced pests, unimportant in their nativehabitats, often reach damaging levels when releasedfrom the control of co-evolved natural enemies. Thebanana weevil appears to fit this pattern. Although thereis some belief that the weevil might reach pest status inparts of Asia, the weevil is generally not considered tobe a serious pest in Asia. Greathead et al. (1986) esti-mated the chances for a successful classical biologicalcontrol programme at 30%.

Therefore, exploration for banana weevil naturalenemies in Asia followed by selection, quaran-tine and release of suitable species could establishan herbivore equilibrium below economic thresh-olds. The first searches for natural enemies in Asiawere undertaken by Muir in 1908 (Froggatt 1925),

Jepson (1914) and Froggatt (1928). They identifiedPlaesius javanus Erichson (Coleoptera:Histeridae),Belonuchus ferrugatus Erichson (Coleoptera: Staphyl-inidae), Leptochirus unicolor Lepeletier (Coleoptera:Staphylinidae), Canthartus sp. (Coleoptera:Cucujidae)and Chrysophila ferruginosa (Wied) (Diptera:Rhagionidae) as being predacious on the bananaweevil and banana stem weevil Odoiporus longicollisOliv. (Coleoptera:Curculionidae). Of these the mostimportant appeared to be P. javanus whose larvaeand adults both attack banana weevil immatures.Later searches revealed the presence of other preda-tors, e.g. Hololepta spp.(Coleoptera:Histeridae) andDactylosternum hydrophiloides MacLeay (Coleoptera:Hydrophilidae) (Table 7a).

The life history of P. javanus has been described byJepson (1914), Froggatt (1928), Weddell (1932) andBarrera & Jimenez (1994). The females place singleeggs under leaves, at the base of plants and on cropresidues. The adults can live up to 14 months, whilethe immature stages last 5–6 months. P. javanus isan opportunistic generalist predator that will feed ona range of prey. In the laboratory, the adults and larvaecan eat up to 8 and 40 banana weevil larvae per day,respectively. This predator is most commonly found in

Table 7. Prospects for classical biological control of the banana weevil C. sordidus: Naturalenemies in area of origin and summary of earlier classical biological control attempts

a. Common natural enemies of banana weevil in Southeast Asia

Coleptera

Histeridae Plaesius javanus ErichsonHyposolenus (Plaesius) laevigatus (Marseul)Hololepta quadridentata (F.)Hololepta spp.

Staphylinidae Belonuchius ferrugatus ErichsonLeptochirus unicolor Lepeletier

Silvanidae Cathartus sp.Hydrophilidae Dactylosternum hydrophiloides MacLeay

Dactylosternum abdominale (F.)DipteraRhagionidae Chrysophila ferruginosus (Wied.)

b. Introductions of natural enemies for the biological control of banana weevil

Insect Attempts Established Location established

Plaesius javanus 27 10 Fiji, JamaicaHyposolenus laevigatus 2 2 Cook Island, DominicaDactylosternum abdominale 1 0D. hydrophiloides 4 2 Australia, JamaicaHololepta quadridentata 7 1 Saint VincentHololepta spp 3 0Chrysophylus ferruginous 1 0Total 45 15

Sources: Gold (1998a), Hasyim & Gold (1999) adapted from Viswanath (1976),Waterhouse & Norris (1987), Geddes & Iles (1991), Waterhouse (1993).


deteriorating banana residues and rarely enters weevilgalleries in living plants.

Between 1913 and 1959, 45 attempts were made tointroduce 8 natural enemies from Asia to other banana-growing regions in the world (Table 7b). P. javanuswas released in Australia, Oceana, Latin America andAfrica. Most commonly, the introductions were donewith small predator consignments and with disregardfor the ecological similarities between source and tar-get sites. In most cases, the natural enemies havefailed to establish following introduction (Hoyt 1957;Greathead et al. 1986; Waterhouse & Norris 1987) or,if established, failed to live up to expectation. Only inFiji and Jamaica has there been any suggestion of evenpartial control (Greathead et al. 1986; Waterhouse &Norris 1987).

Although the lack of success in trying to estab-lish natural enemies introduced from Asia to otherbanana-growing regions in the world is discouraging,additional search efforts would be desirable. Thesesearches might focus on parasitoids (especially of therelatively vulnerable egg stage) which might be host-specific to banana weevil and/or banana stem weevil(Neuenschwander 1988). Such natural enemies tend tobe more effective biological control agents than oppor-tunistic predators such as P. javanus. However, egg par-asitoids are difficult to find and their efficiency may beinfluenced by cultural practices (e.g. mulching) whichwill influence exposure of oviposition sites.

More recently, searches for natural enemies ofbanana weevil were carried out by the InternationalInstitute for Tropical Agriculture (Nigeria) and theResearch Institute for Fruits (Solok, Indonesia) at atotal of 5 sites in Sumatra, Indonesia in 2000 and2001. More than 19,000 eggs and 1500 larvae werecollected in the field and reared in the laboratory. Theeggs were maintained on filter paper in petri dishes,while the larvae were reared on heat-sterilised bananacorm material. A phorid fly, (Megaselia sp.) was rearedfrom several banana weevil larvae, although it is notclear if this was a parasitoid or saprophage. Otherwise,no parasitoids emerged from this material (Abera et al.unpubl. data).

2. Endemic natural enemies

Lists of endemic arthropod natural enemies ofbanana weevil have been provided for Latin America(Mesquita & Alves 1984; Arroyave 1985; Castrillon1991; Londono et al. 1991; Pena & Duncan 1991;Schmitt 1993; Garcia et al. 1994; Sponagel et al. 1995;

Goitia & Cerda 1998; Castrillon 2000), Africa(Koppenhofer 1993b,c, 1994, 1995; Koppenhofer &Schmutter 1993; Koppenhofer et al. 1992, 1995;Tinzaara et al. 1999a) and Asia outside the pre-sumed area of weevil origin (Seshu Reddy et al. 1998;Padmanaban et al. 2001). Cuille (1950), Simmonds(1966), Beccari (1967), and Schmitt (1993) also pro-vide lists of known natural enemies against bananaweevil. Reported natural enemies include nabids,cydnids, capsids, reduviids, mirids, thrips, rhagion-ids, sarcophagids, histerids, carabids, hydrophilids,staphylinids, dermaptera, curculionids, scarabaeids,tenebrionids, and formicids. Vertebrates reported tofeed on banana weevil adults include the giant toadBufo marinus (Dawl 1985), common large arboreallizard Anolis cristatelus (Wolcott 1924) rats, bandi-coots, frogs, birds (Hely et al. 1982) and servals (Goldet al. pers. observ.). Very little information is availableon the efficacy of these natural enemies. Most appearto be of little importance.

Koppenhofer et al. (1992) listed 12 preda-tors of banana weevil in western Kenya. Theseincluded adults of Thyreocephalus interocularis(Eppelsheim) (Coleoptera:Staphylinidae), Hesperiussparsior (Bernhauer) (Coleoptera:Staphylinidae),Charichirus sp. (Coleoptera:Staphylinidae), Histerniloticus Marseul (Coleoptera:Histeridae), Hololeptastriaditera Marseul (Coleoptera:Histeridae),D. abdominale (Fabr.) (Coleoptera:Hydrophilidae),Abacetus optimus Peringuey (Coleoptera:Carabidae), Eutochia pulla Erichson (Coleoptera:Tenebrionidae), Labia curvicauda (Motschulsky)(Coleoptera:Labiidae) and L. borellii Burr (Coleoptera:Labiidae), Euborellia annulipes (Lucas) (Dermaptera:Carcinophoridae), and an unidentified histerid. Theorigin of D. abdominale is unclear. Koppenhofer &Schmutterer (1993) described it as indigenous to EastAfrica, while Koppenhofer et al. (1995) reported that itwas introduced from Malaysia to other banana-growingregions of the world. Waterhouse & Norris (1987) lista single unsuccessful attempt (in Jamaica) with thisspecies.

Koppenhofer (1994, 1995), Koppenhofer &Schmutter (1993), and Koppenhofer et al. (1995) didin depth studies on the bionomics and control potentialof T. interocularis, E. annulipes, adominale. In labo-ratory experiments, these predators variously searchedcorms of living plants and residue pseudostems andcorms (Koppenhofer et al. 1992). Eleven predatorsattacked the banana weevil egg stage, ten attackedthe first two larval instars, nine attacked the third and


fourth instar, while four attacked later stages. The lar-vae of T. interocularis and D. abominalae were alsopredacious on weevil eggs and larvae.

Using high predator densities (i.e. 10–30 adults inplastic containers, 10–200 adults in cages) under exper-imental conditions, D. abdominale reduced weevils byup to 50% in suckers, 39% in stumps, and 40–90% inresidue pseudostems, T. interocularis reduced weevildensities in spent pseudostems by 42%, while theother predators were unimportant (Koppenhofer &Schmutterer 1993; Koppenhofer 1995). However, thenumber of natural enemies used in these experimentswell exceeded field densities suggesting that the impactof these predators in banana stands is likely to be lim-ited (Koppenhofer & Schmutterer 1993). Field evalu-ation of natural enemy efficacy was not carried out forany of these species.

Hargreaves (1940) was the first to suggest that antsmight have potential as biological control agents ofbanana weevil in Africa, although no studies were con-ducted. During the 1970s, Cuban researchers began abiological control programme using the myrmicine antPheidole megacephala (Hymenoptera:Formicidae)against sweet potato weevils (Perfecto 1994). Based onanecdotal reports that plantain stands lasted 15 yearswithout the application of insecticides and only2–3 years where chemicals were applied, Roche (1975)deduced the presence of effective natural enemiesand suggested that Tetramorium guineense (Mayr)might be capable of suppressing banana weevil pop-ulations. T. guineense was observed to nest in leavesand galleries and to keep plants free of weevils andfrass. A programme was then initiated employingthe use of T. guineense and P. megacephala againstthe banana weevil. The ants have been observed toenter crop residues and remove eggs and larvae(S. Rodriguez pers. comm.). According to Perfecto &Castineiras (1998), P. megacephala can also reduceweevil oviposition when they nest near the plant roots.

Roche & Abreu (1982, 1983) began the propa-gation and dissemination of T. guineense colonies.Colonies with up to 22 queens and 62,000 workersand immatures were collected and liberated in newfields (Perfecto & Castineiras 1998). Ant establish-ment was followed by the rapid appearance of newcolonies. At the onset of one trial, weevil trap catcheswere greater than 10/mat. Liberation of ants on 8%and 50% of the mats provided total field coverage in6 and 2 months, respectively (Roche & Abreu 1983).Eighteen months later, the high colony release rate hadreduced weevil populations by 65%, while trap catches

were 56% lower in the low release rate. By compar-ison, pesticide applications reduced trap captures by79%. Based on these results, Roche & Abreu (1983)recommended releasing ants on 25–30% of the areafor ‘complete control’ in 3–4 months.

The control potential of myrmicine ants has alsobeen demonstrated by Castineiras & Ponce (1991).They released 9 and 15 P. megacephala colonies/hainto plantain plots (separated by 200 m alleys) 6 monthsafter planting. During the first crop cycle, weevil trapcaptures, and damage indices (CI of Vilardebo (1973))were similar in plots where ants had been released,in carbofuran treated plots and controls. In the secondcycle, ants reduced weevil trap captures by 54–69%and damage by 64–66%, with a corresponding yieldincrease of 15–22%. The level of control of ants wassimilar to that of the pesticide.

Castineiras (1982) studied the diurnal activity andseasonality of P. megacephala and found the ant for-aged throughout the day with greater activity duringdaytime in winter and greater and during night time insummer. Similarly, Roche & Perez (1985) found con-tinuous activity of T. guineense with reduced activity athigh temperatures and greater activity with increasingrelative humidity. Bendicho & Gonzalez (1986) foundlower levels of control with T. guineense in the dry sea-son, even though ant numbers were higher at that time.

The Cubans have since developed an integrated con-trol strategy against banana weevil using T. guineense,P. megacephala, and the entomopathogen B. bassiana(S. Rodriguez pers. comm.). The ants alone have beenreported to provide 60–70% control (Perfecto 1994),although the studies have not been well documentedand only few data have been published. Farmers havereportedly seen the value of these predacious antsand often place molasses and kitchen scraps aroundtheir banana plants to encourage the ants (Perfecto &Castineiras 1998). The use of pesticides has been pro-hibited in areas where biological control programmesagainst banana weevil are in operation.

Based on the results gained in Cuba, Greathead(1986) suggested that ants might have potentialfor biological control of banana weevil in Africa.Waterhouse & Norris (1987) also recognised the poten-tial of ants for weevil control in Asia and the Pacificand proposed that they might be evaluated in areas(e.g. Solomon Islands, Papua New Guinea) where theweevil is not important.

Walker & Dietz (1979) found four species ofTetramorium (including T. guineensee) and twospecies of Pheidole including (P. megacephala) in the


Cook Islands. Varela (1993) found 40 species of antsin a survey of banana stands in four areas in Kageradistrict, Tanzania. Pheidole was the most abundantgenus with P. megacephala the dominant species. Sheobserved P. megacephala nesting on the ground andin leaf sheaths and found in large numbers in tunnels.In Uganda, Gold & Nemeye (unpubl. data) surveyedbanana stands in 5 sites and found 35 species of antsincluding T. sericeiventre, P. megacephala, and fiveother species of Pheidole.

Most of the literature is unclear whether these antswill enter galleries in living plants (which are ordinar-ily filled with latex) or only in crop residues. However,Bendicho & Gonzalez (1986) noted that the smallsize of T. guineense allows penetration into larval gal-leries. In one laboratory experiment, Bendicho (1987)inserted banana weevil larvae into planted corms andlater observed T. guineense workers enter the galleriesand remove larvae. Abera (pers. comm.) has observedPheidole spp. removing eggs and larvae from pseu-dostems in laboratory studies. Gold (pers. observ.) alsoobserved small, unidentified (probably myrmicine)ants in weevil galleries in plants at the time of harvest.

In Venezuela, Goitia & Cerda (1998) found15 species of ants (eight myrmicinae, one pseudomyr-micine, two dolochoderinae, three ponerilnae, and oneecitononae) in a 5-year-old banana plantation. Themost common of these were Azteca foreli, Ectatommaruidum, Wasmannia auropunctata, and Odontomachusbaueri. Of these, W. auropunctata and E. ruidum wereconsidered potential predators of the weevil. However,it was not confirmed if either of these species do, infact, predate on weevil immatures.

Traore (1995) surveyed plantain systems in Benin,Cote d’Ivoire and Nigeria for possible egg parasitoidsusing yellow pan traps. He found a wide range of eggparasitoids belonging to 12 genera, of which MymarCurtis, Lymaenon Hal, and Anagrus Haliday were mostcommon. However, he was unable to find any indi-cations of parasitism of banana weevil eggs collectedfrom plants, placed in the field in infested suckers,or attached to yellow cards. Similarly, Koppenhofer(1993c) and Abera (unpubl. data) were unable to detectegg parasitism in Kenya and Uganda, respectively.

3. Secondary host association

Neuenschwander (1988) suggested that natural ene-mies of closely related hosts offer the promise forefficient secondary associations with banana weevil.Traore (1995) investigated the possible use of the

mymarid egg parasitoid Anaphes victus Huber againstbanana weevil in Benin. A. victus is an important par-asitoid of weevil eggs in the Americas. This parasitoidwas selected for study because it searches near the soillevel, is habitat rather than species specific and becauseit effectively suppresses populations of carrot weevil(Listronotus oregonensis (LeConte)) (Boivin 1993).Traore’s study tested two A. victus biotypes (Quebecand Texas) reared from carrot weevil eggs.

In the laboratory, A. victus readily accepted bananaweevil eggs with 60% parasitism by the Quebec bio-type and 35% by the Texas biotype. However, par-asitoid emergence was negligible (2% and 0% fromthe Quebec and Texas biotypes, respectively) (Traore1995). In contrast, A. victus immatures successfullyemerged from water hyacinth weevil, Neochetinaeichhorniae Warner, demonstrating that the parasitoidcould successfully complete its development withina new host. Traore (1995) attributed these disparateresults to differences in host egg size. Banana weevileggs were considerably larger than those of carrotor water hyacinth weevils. Larvae of A. victus failedto consume all of the banana weevil egg contentswith decomposition of unconsumed material contribut-ing to pupal failure. Most of the few parasitoids thatsuccessfully reached the adult stage then failed toemerge through the relatively thicker chorion of bananaweevil eggs.

XIII. Microbial Control

Research on microbial control of banana weevilis still in its early stages. Microbial agentstested against the weevil include entomopathogenicfungi (e.g. B. bassiana and M. anisopliae), ento-mopathogenic nematodes (e.g. Steinernema spp. andHeterorhabditis spp.) and endophytes (e.g. non-pathogenic Fusarium spp.). Entomopathogenic fungiand nematodes are most often used to kill adult weevils,while endophytes target the immature stages. Althougha number of strains have shown promise in the labo-ratory and in preliminary field studies, efficient andeconomically viable mass production and delivery sys-tems still need to be developed, while the performanceof microbial control agents against banana weevilunder different agro-ecological conditions is not wellunderstood.

Epizootics of entomopathogenic fungi or nema-todes in nature are uncommon, while natural infectionrates of banana weevil tend to be quite low. Only in a


few sites have entomopathogens been reported to estab-lish following applications in banana fields. Withoutadequate establishment, entomopathogens will requirerepeated applications as a biopesticide. This will entailcontinued production, distribution and storage coststhat will be passed on to the farmer.

1. Entomopathogenic fungi

a. Research protocols and strain selectionNankinga (1994, 1997) noted that species in the ‘fungiimperfecti’ may survive as saprophytes making thembetter candidates as biological control agents thanfungi that are obligate parasites. Two genera withinthis group, Beauveria and Metarhizium, are widelydistributed and have been reported from hundreds ofinsect hosts. The most common species are B. bassiana,B. brongniartii, and M. anisopliae.

B. bassiana and M. anisopliae have gained consider-able attention as biological control agents for weevilsand other agricultural pests (Ferron 1981). These areespecially important for controlling cryptic insects,such as banana weevil, which are not accessible toarthropod natural enemies. The ability of the ento-mopathogen to survive and infect soil-dwelling insectsis a primary determinant of efficacy under field condi-tions. Pathogen viability can range from days to years,depending on ecological conditions and the applica-tion method used. Different strains of B. bassiana andM. anisopliae have distinct ecological requirements(e.g. temperature, humidity, and soil pH) which deter-mine the environmental conditions under which theyare most effective.

The conidia of Beauveria and Metarhizium enter theinsect through its spiracles or digestive system or byproducing extracellular proteolytic, chitinolytic, andlipolytic enzymes which facilitate penetration throughthe insect’s cuticle (Nankinga 1999). B. bassiana canalso adhere to the cuticle and penetrate the integumentthrough a germ tube (Godonou 1999). The fungi cankill the insect through direct attack on the insect’s nutri-ents or through toxic metabolites (Nankinga 1997).For example, in banana weevil, Kaaya et al. (1993)observed that chains of B. bassiana or M. anisopliaehyphae invade the haemocoel and muscle tissues anddestroy tracheal taenidia and fat bodies. Dead insectskept in moist environment quickly developed surfacegrowth of mycelia. B. bassiana can invade the haemo-coel where it produces a toxin, beauvericin, that reducescompetition with bacteria and weakens the immunesystem (Hamill et al. 1969). Strain virulence is often

related to toxin production (Ferron 1981). After killingtheir hosts, the fungus can live saprophytically.

Pathogenicity studies on banana weevil havebeen done in disperse locations (Africa, Australia,Latin America) against populations of weevils thatmay represent distinct biotypes (c.f. Ochieng 2002).Pathogenicity of fungal isolates may be affected by:(1) the source of the isolate (Brenes & Carballo 1994);(2) the method of culturing (Altre & Vandenberg2001); (3) spore dose (Nankinga 1994; Godonou 1999);(4) temperature and relative humidity (Fargues & Luz2000; Arthurs & Thomas 2001); (5) formulation andmode of application (Nankinga 1994; Godonou 1999;Godonou et al. 2000). Under field conditions, fun-gal efficacy may also be affected by factors influenc-ing soil moisture (e.g. soil type, precipitation patterns,mulch). For example, Nankinga (1999) suggested thatmulching might prolong the life of the fungus, but alsonoted that an increase in soil moisture might leadto more rapid degradation of B. bassiana by othersoil microorganisms. Pena et al. (1993) and Traore(1995) found higher rates of infected weevils whenB. bassiana was applied to sterilised soil than whenapplied to non-sterile soil, further suggesting theantagonistic action of other soil organisms. Nankinga(1999) suggests that microbial degradation of ento-mopathogens may be more pronounced in high organicsoils than in clay soils. In general, more inoculum isrequired for the control of soil-borne insects (Godonou1999).

Research protocols for the development of a micro-bial control programme of banana weevil includea series of steps starting at isolation and screeningof candidate strains to the development of econom-ical and effective field delivery systems (Godonou1999). Efficient mass production systems are criticalfor programmes that will depend on augmentativeor inundative releases. Candidate strains of micro-bial agents are often selected from existing collec-tions, from dead weevils found in the field or byGalleria bait methods to obtain fungi in soils inbanana plantations (Castineiras et al. 1990; Brenes &Carballo 1994; Nankinga 1994, 1999). For exam-ple, B. bassiana strains screened against bananaweevil have been isolated from hemiptera, lepidoptera,coleoptera (including banana weevil and other weevils)and hymenoptera (i.e. ants). Although B. bassianaisolates are usually most pathogenic to the orig-inal or related hosts (Nankinga 1999), some ofthe best performing strains had been isolated fromnon-coleopterans (Brenes & Carballo 1994).


Strains effecting high kill rates in the laboratory needto be characterised to determine sporulation rates, theirpotential for mass production on a range of substrates,and spore viability following storage and performanceunder different ecological conditions. Further testingwill then evaluate candidate strains for field effi-cacy at different fungal concentrations, under differ-ent formulations and for a range of delivery systems.Ultimately, the capacity to deliver entomopathogensto farmers, costs of application, and the level of con-trol will determine the feasibility of a microbial controlmethod.

Most research on entomopathogens and bananaweevil has focused on levels of adult mortality in thelaboratory, in pot trials and/or in small pilot field stud-ies. While many of these studies show promise, littlework has been done in larger trials or at the farmer levelto show the true potential of microbial control agents.This makes it difficult to interpret and integrate thebody of literature that has been developed on the effi-cacy of different strains and application formulationsconcentrations and rates.

Entomopathogenic fungi have been tested againstbanana weevil since the 1970s (Ayala & Monzon1977; Delattre & Jean-Bart 1978). Since then, numer-ous laboratory studies conducted in many differentbanana-growing regions have demonstrated high lev-els of mortality to a large number of strains (Table 8).Additional research has been conducted on strainselection, mass production on a range of substrates(e.g. maize, rice), spore viability and storage, formula-tions (powders, water solutions, mineral oils) and shelflife, doses, application methods, and mortality ratesafter varying time intervals. Few studies have addresseddelivery systems and efficacy under field conditions.The use of entomopathogens against banana weevil hasbeen reviewed by Nankinga et al. (1999).

b. Natural infection in banana fieldsIn Brazil, Mesquita et al. (1981) assessed field infec-tion levels of both banana weevil and Metamasiushemipterus collected in pseudostem traps over9 months. Monthly infection rates were not differenti-ated by species and ranged from 1–7% with an overallmean of 3%. De Souza et al. (1981) found field infec-tion to average 1–2%, with a peak of 8%. Infectionrates were negatively correlated (−0.35 to −0.44) withweevil population levels. Batista Filho et al. (1992)found 9% infection of banana weevils by B. amorphain pseudostem traps in a Prata stand and no infection inan adjacent Nanica plot. In Cuba, Gomes (1985) found

<1% field-collected weevils infected by B. bassiana,while in Colombian surveys, Van den Enden & Garcia(1984) found only five infected adults and two infectedimmatures.

In Florida, Pena et al. (1993) reported 6% naturalinfection of banana weevils by B. bassiana in one study.In a second field, weekly assessment of weevils in pseu-dostem traps showed infection rates to range 4–34%;mortality of >10% occurred in 4 of 38 sampling peri-ods (Pena et al. 1995). Many additional dead weevilsinfected with B. bassiana were found adhering to theunderside of fallen banana pseudostems and corms.In this study, infection rates tended to rise followingincreases in weevil population levels.

In Uganda, Nankinga (1994) isolated B. bassianaor M. anisopliae (using Galleria larvae) from 29 and3 samples, respectively, out of 37 soil samples takenfrom banana stands in total of 24 sites. However, Gold(pers. observ.) and Nankinga (unpubl. data) found natu-ral infection to be 0–3% among hundreds of thousandsof weevils collected in pseuodstem traps and amongtens of thousands of these weevils maintained in thelaboratory.

In a survey of the Department of Risaralda,Colombia, Castrillon (2000) found B. bassiana in 6of 9 municipalities with an incidence of 0–11% (mean4%) infection of weevils collected in pseudostem trapsand maintained for 15 days in the laboratory.

c. Spore production, formulations, and viabilityMethods for mass production and delivery of ento-mopathogens have been reviewed by Nankinga (1999)and Godonou (1999). Production systems includedliquid fermentation, solid substrates, and diphasicmethods. Substrates offering large surface areas forfungal sporulation are normally preferable. In addi-tion to choice of substrate, moisture content, balanceof nutrients, pH, and aeration may also affect conidialor spore production (Godonou 1999). Pathogen appli-cations can be made in dry state using solid substratecarriers, as wettable powders, dusts, baits or granules,or in oil- or water-based liquid sprays. Formulationsare also important in stabilising the pathogen, improv-ing efficacy in field and providing an economic andeasily usable form of active ingredient with longshelf life (Godonou 1999). Nankinga (1999) foundmaize to be the best solid substrate as it was associ-ated with high sporulation, low contamination, >95%germination and 60–100% infectivity of weevils in14 days. Oil may provide greater adhesiveness to cuti-cle, increase the number of conidia reaching the insect’s


intersegmental membranes, enhance protection againstultraviolet light and desiccation, and cause higher mor-tality at lower doses (Prior et al. 1988; Nankinga 1999).However, oil-based formulations are more costly thanother formulations.

Batista Filho et al. (1987) reported 75% viabilityof B. bassiana and M. anisopliae spores inoculatedinto solid rice and 85% viability in liquid substrates.Nankinga (1994) compared spore production amongdifferent B. bassiana isolates and found that some didbetter at ambient temperatures, while others performedbetter at higher temperatures. She also found sporeviability for up to 2 years. Ferreira (1995) estimatedspore viability for four strains of B. bassiana to be75–95%. Godonou (1999) and Godonou et al. (2000)used the number and weight of conidia per unit sub-strate and conidial viability, to complement virulencelevels, in selection of candidate strains for further test-ing. Conidia obtained from rice and oil palm kernelcake substrates displayed 98% viability.

d. Mortality rates for adultsNumerous strains of entomopathogens have beenscreened against banana weevil in the Americas andAfrica, employing a range of formulations, sporeconcentrations, and application methods. Mortalityof weevils exposed to many strains often reached90–100% (Table 8).

In Cuba, Ayala & Monzon (1977) found 50–70%mortality of weevils 33 days after being released incages at the base of banana mats treated with 4, 8, 12,or 16 g of B. bassiana (2×105 spores/mg). Castineiraset al. (1990) evaluated 17 strains of B. bassiana (mostlyfrom lepidoptera) and 11 strains of M. anisopliae(including three from banana weevil) for efficacyagainst banana weevil following 1 min dips in watersuspensions (2×108 spores/ml). After 30 days, mortal-ity ranged 15–58% and only one strain of each specieseffected mortality of >50%.

In Guadeloupe, Delattre & Jean-Bart (1978)screened six strains of B. bassiana, two strains ofB. brongniartii, five strains of M. anisopliae, and onestrain of Nomuraea rileyi against banana weevil in thelaboratory using impregnated filter paper. Three strainseach of B. bassiana and M. anisopliae caused mortalityreaching 60–100% after 90 days. The remaining strainsof B. bassiana and M. anisopliae and the tested isolatesof B. brongniartii and N. rileyi were all ineffective. Incontainer experiments, spores or conidia applied to theloamy soil resulted in 0–15% mortality, while sporesapplied to clay caused 52–54% weevil mortality.

In the West Indies, Khan & Gangapersad (2001)found LD50 and LT50 values of 4.57 × 107 spores/mland 10 days, respectively for B. bassiana, 5.13 ×107 spores/ml and 21 days for M. anisopliae, and4.92 × 108 spores/ml and 32 days for M. flavoviridae.

In Brazil, Gomes (1985) applied a single strain ofB. bassiana in a spore suspension (powder) at a con-centration of (2 × 109 spores/mg). Direct immersionof weevils resulted in mortality of 22%, while applica-tions to pseudostem traps and soil resulted in mortalityof 16% and 34%, respectively.

In laboratory studies, Batista Filho et al. (1987)reported weevil mortality to B. bassiana andM. anisopliae to be 85% and 93%, respectively, in rice-based substrates and 97% and 56%, respectively, inliquid formulations. After 16 days, Batista Filho et al.(1994) found 38% and 78–100% mortality of weevilsin pseudostem traps treated with B. bassiana alone andas a homogenised rice paste + mineral oil formula-tion. In another experiment, Batista Filho et al. (1995a)observed 70% and 98% mortality when weevils wereexposed to rice substrate and mineral oil formulationsof B. bassiana, respectively. The proportion of weevilsshowing fungal symptoms was similar (60%) in thetwo formulations. However, a mineral-based formula-tion caused more rapid mortality (e.g. 88% at 8 days)than the rice substrate (14%).

Busoli et al. (1989) tested two strains each ofB. bassiana (isolated from a pyralid and a scarab) andof M. anisopliae (isolated from a scarab and a cer-copid). The fungi were produced on rice and appliedtopically as a powder with 1000 or 2000 spores perinsect. The two doses of B. bassiana caused mortalityof 32% and 80%, respectively at 10 days and 61% and99% mortality at 33 days. In comparison, the two dosesof M. anisopliae caused mortality of 15% and 68%,respectively at 10 days and 47% and 79% mortality at33 days.

In Florida, Pena et al. (1993) tested three isolatesof B. bassiana against banana weevil. Mortality of>40% was achieved with 107 or 108 spores/g soil.Virulence was greater on sterile soils than on non-sterile soils. Water-saturated soils had significantlyhigher levels of weevil mortality (>35%) than drysoils (10%).

In Costa Rica, Brenes & Carballo (1994) screened24 isolates of B. bassiana (from hemiptera, lepidoptera,ants and other weevils) by shaking the insects in coni-dial powder. The six most promising isolates wereselected for further testing. Mortality of weevils dippedin water suspensions containing 1 × 109 spores/ml

Tabl

e8.

Test

ing

and

scre

enin

gof

ento

mop

atho

gens

agai

nstb

anan

aw

eevi

l:Su

mm

ary

ofre

sear

chm

etho

dsan

dw

eevi

ladu

ltm

orta

lity

leve

ls

Cou

ntry

Spec

ies

Stra

ins

Form

ulat

ion

Spor

es/m

gor

ml

App

licat

ion

Tim

eLT

50M

orta

lity

Ref

eren

cepe

riod

(%)

Ben

inB

.bas

sian

a1

Pean

utoi

l1.

1×

103–1

.1×

108

Filte

rpa

per

214–

96T

raor

e(1

995)

1Pe

anut

oil

1.1

×10

3–1

.1×

108

Soil

2124

–67

Tra

ore

(199

5)1

Pean

utoi

l1.

1×

103–1

.1×

108

Filte

rpa

per

218–

84T

raor

e(1

995)

1Pe

anut

oil

1.1

×10

3–1

.1×

108

Soil

2140

–73

Tra

ore

(199

5)B

razi

lB

.bas

sian

a1

Pow

der

85B

atis

taFi

lho

etal

.(19

87)

1L

iqui

d97

Bat

ista

Filh

oet

al.(

1987

)5

Spor

ecu

lture

1×

108

Topi

cal

1–15

2–40

Bat

ista

Filh

oet

al.(

1991

)1

Ric

epa

ste

1.8

×10

9T

raps

1638

Bat

ista

Filh

oet

al.(

1994

)1

Min

eral

oil

1.8

×10

9T

raps

1638

–100

Bat

ista

Filh

oet

al.(

1994

)1

Min

eral

oil

1.8

×10

9T

raps

1653

–78

Bat

ista

Filh

oet

al.(

1994

)1

Ric

ecu

lture

5×

106

Tra

ps4–

2070

Bat

ista

Filh

oet

al.(

1995

b)1

3%M

in.o

il5

×10

6T

raps

4–20

98B

atis

taFi

lho

etal

.(19

95a)

2Po

wde

r1–

2000

spor

esTo

pica

l33

61–9

9B

usol

i(19

89)

2R

ice

past

e1

×10

8T

raps

2090

–92

Ferr

eira

(199

5)1

Pow

der

2×

109

Imm

ersi

on22

Gom

es(1

995)

1Po

wde

r2

×10

9T

raps

16G

omes

(198

5)1

Pow

der

2×

109

Soil

34G

omes

(198

5)4

Wat

er5

×10

9D

ispe

rsio

n4–

3673

–100

Mes

quita

(198

8)4

Wat

er5

×10

9So

il4–

3663

–67

Mes

quita

(198

8)4

Wat

er5

×10

9T

raps

4–36

56–1

00M

esqu

ita(1

988)

1W

ater

1×

107

Dis

pers

ion

1570

–90

Soar

eset

al.(

1980

)1

Wat

er1

×10

7T

raps

2010

0So

ares

etal

.(19

80)

M.a

niso

plia

e1

Ric

epa

ste

93B

atis

taFi

lho

etal

.(19

87)

1L

iqui

d56

Bat

ista

Filh

oet

al.(

1987

)2

Pow

der

1–20

00sp

ores

Topi

cal

3347

–79

Bus

olie

tal.

(198

9)1

Wat

er5

×10

9D

ispe

rsio

n4–

3640

Mes

quita

(198

8)1

Wat

er5

×10

9So

il4–

3666

Mes

quita

(198

8)1

Wat

er5

×10

9T

raps

4–36

40M

esqu

ita(1

988)

Col

ombi

aB

.bas

sian

a1

Ric

epa

ste

Tra

ps8

39G

arci

aet

al.(

1994

)C

osta

Ric

aB

.bas

sian

a6

Wat

er1

×10

9D

ips

7–10

73–1

00B

rene

s&

Car

ballo

(199

4)1

Wat

er4

×10

5–4

×10

9D

ips

7–11

63–

98B

rene

s&

Car

ballo

(199

4)6

Wat

er2.

67×

10D

ips

9–19

50–9

8B

rene

s&

Car

ballo

(199

4)1

Ric

esu

bstr

ate

5.8

×10

10T

raps

10–1

131

–33

Car

ballo

&de

Lop

ez(1

994)

1Po

wde

r5.

8×

1010

Tra

ps11

–13

33–6

3C

arba

llo&

deL

opez

(199

4)1

10–2

0%oi

l5

×10

8D

ispe

rsio

n6

100

Cab

allo

(199

8)1

15%

oil

1×

107–5

×10

8D

ispe

rsio

n8–

3010

–97

Cab

allo

(199

8)

515

%oi

l1

×10

8D

ispe

rsio

n15

3–8

65–9

5C

ontr

eras

(199

6)1

15%

oil

5×

108

Tra

ps61

Con

trer

as(1

996)

1R

ice

subs

trat

e2.

75×

109/g

rice

Tra

ps85

Con

trer

as(1

996)

Cub

aB

.bas

sian

a1

Pow

der

2×

105

Soil

3350

–70

Aya

la&

Mon

zon

(197

7)17

Wat

er2

×10

8D

ips

300–

56C

astin

eira

set

al.(

1990

)M

.ani

sopl

iae

11W

ater

2×

108

Dip

s30

0–58

Cas

tinei

ras

etal

.(19

90)

Gha

naB

.bas

sian

a1

Wat

erT

raps

59G

odon

ou(1

999)

1W

ater

Soil

(pot

s)24

–62

God

onou

(199

9)1

Pow

der

Suck

ers

53–8

1G

odon

ou(1

999)

Gua

delo

upe

B.b

assi

ana

6W

ater

2×

106–

2×

108

Filte

rpa

per

0–90

10–1

00D

elat

tre

&Je

an-B

art(

1978

)B

.bas

sian

a1

Wat

er1

×10

7So

il60

0–15

Del

attr

e&

Jean

-Bar

t(19

78)

B.b

assi

ana

1W

ater

1×

107

Cla

y60

52–5

4D

elat

tre

&Je

an-B

art(

1978

)B

.bro

ngni

arti

i2

Wat

er2

×10

6–2

×10

8Fi

lter

pape

r0–

9010

–20

Del

attr

e&

Jean

-Bar

t(19

78)

M.a

niso

plia

e5

Wat

er2

×10

6–2

×10

8Fi

lter

pape

r0–

9010

–70

Del

attr

e&

Jean

-Bar

t(19

78)

N.r

iley

i1

Wat

er2

×10

6–2

×10

8Fi

lter

pape

r0–

9010

Del

attr

e&

Jean

-Bar

t(19

78)

Ken

yaB

.bas

sian

a4

Spor

ecu

lture

sTo

pica

l9

3–4

Lar

v.90

–100

Kaa

yaet

al.(

1993

)4

Spor

ecu

lture

sTo

pica

l35

8–22

Adu

lt60

–98

Kaa

yaet

al.(

1993

)M

.ani

sopl

iae

1Sp

ore

cultu

res

Topi

cal

94

Lar

vae

98K

aaya

etal

.(19

93)

1Sp

ore

cultu

res

Topi

cal

35A

dult

28K

aaya

etal

.(19

93)

Sout

hA

fric

aB

.bas

sian

a1

Wat

er1.

33×

109

Topi

cal

3710

0Sc

hoem

an&

Scho

eman

(199

9)U

gand

aB

.bas

sian

a6

Spor

ecu

lture

sTo

pica

l5–

2142

–98

Nan

king

a(1

994)

6W

ater

2.28

×10

8To

pica

l5–

1721

–100

Nan

king

a(1

994)

3W

ater

3.35

×10

7To

pica

l7–

3293

–96

Nan

king

a(1

994)

3W

ater

3.35

×10

6To

pica

l7–

3260

–69

Nan

king

a(1

994)

3W

ater

3.35

×10

5To

pica

l7–

3222

–37

Nan

king

a(1

994)

3W

ater

3.35

×10

4To

pica

l7–

328–

19N

anki

nga

(199

4)3

Wat

er1.

12×

107

Dis

pers

ion

2856

–62

Nan

king

a(1

994)

3W

ater

1.12

×10

7Im

mer

sion

2858

–69

Nan

king

a(1

994)

3W

ater

1.12

×10

7So

il28

6–10

Nan

king

a(1

994)

3W

ater

1.12

×10

7T

raps

2810

–11

Nan

king

a(1

994)

15W

ater

6×

109

Imm

ersi

on30

2.5-

100

Nan

king

aet

al.(

1996

)B

.bro

ngna

tii

1W

ater

6×

109

Imm

ersi

on30

85N

anki

nga

etal

.(19

96)

B.s

teph

anod

eris

1W

ater

6×

109

Imm

ersi

on30

2.50

Nan

king

aet

al.(

1996

)M

.ani

sopl

iae

1Sp

ore

cultu

res

Topi

cal

5–21

40N

anki

nga

(199

4)1

Wat

er2.

28×

108

Topi

cal

5–17

30N

anki

nga

(199

4)15

Wat

er6

×10

9Im

mer

sion

3032

.5–9

7.5

Nan

king

aet

al.(

1996

)U

nite

dSt

ates

B.b

assi

ana

3W

ater

10–1

08So

il1–

65Pe

naet

al.(

1993

)1

Wat

er1

×10

2–1

×10

6So

il6–

47Pe

naet

al.(

1993

)


ranged 73–100% with a LT50 of 7–10 days. Usinga range of spore concentrations, an CL90 of 2.67 ×109 spores/ml was calculated for the most promis-ing isolate. Carballo & de Lopez (1994) then found31–63% adult mortality when B. bassiana conidialpowder or spores on rice substrate were applied topseudostem traps.

Contreras (1996) screened five strains of B. bassianain the laboratory in oil-based formulations and found65–95% weevil mortality in 15 days, with an LT50 of2.5–8 days. Carballo (1998) tested water-based and oil-based formulations of B. bassiana. Oil formulations of>20% without fungi caused high levels of mortalityin the weevil, while weevil mortality was negligiblein solutions with 10% oil. Using a 15% oil solution,Carballo (1998) found mortality to range from 10% at1 × 107 spores/ml to 97% at 5 × 108/ml.

In Colombia, Garcia et al. (1994) applied a ricepaste formulation of B. bassiana to pseudostem trapsbiweekly for a 10-month period. Weevil infection,assessed 8 days after application, averaged 39%.

In Kenya, Kaaya et al. (1993) reported four strainsof B. bassiana and one strain of M. anisoplae tocause 90–100% mortality of third-instar larvae. TheB. bassiana isolates killed 60–98% of adult weevilswith LT50 ranging from 8 to 25 days; in contrast, theM. anisoplae killed only 28% of the adults.

In Benin, Traore (1995) found 50% adult mortal-ity in the laboratory at 1.1 × 107 spores/ml for oneexotic strain each of B. bassiana and M. anisopliae.In contrast, soil applications required doses of 1.5 ×108 spores/ml for M. anisopliae and 2.9×108 spores/mlfor B. bassiana to achieve 50% mortality.

In Ghana, Godonou (1999) evaluated adult weevilmortality following applications of different formula-tions of B. bassiana to pots containing plantain swordsuckers or by coating suckers. A groundnut oil pluskerosene formulation and conidial powder induced thehighest rates of mortality (often >80%) immediatelyafter application. However, a formulation utilising oilpalm kernel cake also induced substantial of mortality,while enhancing fungal multiplication and displayingthe greatest level of field persistence. Oil palm kernelcake had the added advantage of being a readily avail-able waste product, while other widely tested substrates(e.g. groundnut, rice, maize) are valuable food crops.

In Uganda, Nankinga (1994) allowed weevils towalk on PDA cultures and found five isolates ofB. bassiana produced >96% mortality after 21 days,while one B. bassiana and one M. anisopliae isolateeach caused only 40% mortality. Topical applications

of the same isolates in water suspensions produced73–100% mortality for five B. bassiana isolates,22% for one B. bassiana isolate, and 30% for theM. anisopliae isolate. Nankinga (1994) also found mor-tality rates to be directly related to spore dose forthree strains of B. bassiana. Higher doses killed almostall weevils, while females were more susceptible thanmales to lower doses of the pathogen. Topical applica-tions by dispersion or immersion caused much higherrates of mortality than spraying pathogen solutions onto soil or pseudostem traps.

In a followup study, Nankinga et al. (1996)screened 15 strains of B. bassiana, one strain eachof B. brongnatii and B. stephanoderis, nine strains ofM. anisopliae, and two strains of M. flavoviride using anaqueous solution with 6×109 spores/ml. After 30 days,14 B. bassiana strains averaged 87% infection (onestrain was ineffective), M. anisopliae strains averaged79% infection, M. flavoviride strains averaged 16%infection, the single strain of B. brongnatii caused 85%infection and the single strain of B. stephoderis killedonly 3% of the weevils.

Nankinga (1999) evaluated a further 31 iso-lates of B. bassiana, 17 of M. anisopliae, two ofM. flavoviride, and one of B. brongniartii. Eighteenisolates gave >70% mortality when weevils wereexposed to 3-week-old sporulating cultures. Whenweevils were exposed to 3–4 ml spore suspension with3 × 1011 spores/ml for 2 h, 22 isolates caused 70–90%mortality. However, when weevils were inoculated with1 ml of a water suspension of the same dose, no isolategave more than 60% mortality. Nankinga (1999) thenselected candidate isolates on the basis of pathogenicitytowards the weevil and growth and sporulation rates.

Although it is difficult to compare the results ofdifferent studies because of the wide range of con-ditions and methods used, several conclusions can bereached. There was wide variability in the efficacy ofdifferent strains in killing banana weevils. The mosteffective strains were capable of causing high mortal-ity in the laboratory at lower spore concentrations andin shorter periods of time. In general, promising iso-lates of B. bassiana were more effective than those ofM. anisopliae (Delattre & Jean-Bart 1978; Batista Filhoet al. 1987; Mesquita 1988; Busoli et al. 1989; Kaayaet al. 1993; Nankinga 1994). However, Castineiras &Ponce (1991) found a mean mortality of 14% for 17 iso-lates of B. bassiana compared 27% for 11 isolates ofM. anisopliae, while Traore (1995) found a single iso-late of M. anisopliae to be more virulent than his isolateof B. bassiana.


Results from several studies (Delattre & Jean-Bart1978; Pena & Duncan 1991; Kaaya et al. 1993;Nankinga & Ogenga-Latigo 1996) suggest that indige-nous isolates might perform better than exotic strains.Topical application tended to produce higher lev-els of mortality than fungal applications to substrates(e.g. soil, traps) where weevils reside. However, fielddelivery systems will have to rely on applicationsto substrates or in areas where weevils might beaggregated by semiochemicals.

e. Mortality to immaturesVan Enden & Garcia (1984), Kaaya et al. (1993), Penaet al. (1993), Nankinga (1994, 1999), and Godonou(1999) reported mortality of weevil immatures toentomopathogens. During field surveys in Colombia,Van Enden & Garcia (1984) observed a single larva andone pupa infected by B. bassiana, suggesting that thefungus can enter banana plants but that effects on imma-ture populations may be minimal. In the laboratory,Kaaya et al. (1993) found >90% mortality of third-instar banana weevil larvae within 9 days of exposurefor each of the four B. bassiana and one M. anisopliaeisolates tested. LT50 times ranged from 3 to 4 days.Assays of the same isolates on adults produced lowerlevels of mortality and LT50 times of 8–22 days.

Pena et al. (1993) allowed weevil ovipositionon suckers that were then immersed in aqueousB. bassiana suspensions and planted in pots. The per-centage of larvae infected with B. bassiana in treatedplants was 43% after 11 days and 13% after 26 days.At 26 days after treatment, larval mortality was 3% incontrols and 13% after 26 days. Pena (unpubl. data)found that B. bassiana injected into a plant could moveabout 30 cm.

In Uganda, dusting suckers with B. bassiana sporesresulted in infection of 41% of the eggs and 19% offirst-instar larvae, while planting suckers in soil treatedwith B. bassiana also led to a low level of larval infec-tion (Nankinga 1994, 1997). In Ghana, soaking of cormand pseudostem pieces in water formulations resultedin up to 46% egg failure and 27% larval mortality, com-pared to 5% and 1% in controls, respectively (Godonou1999).

Kaaya et al. (1993) isolated the bacteria Serratiamaraescens from dead larvae in a colony of bananaweevil. Third-instar larvae were found to be verysusceptible (mortality >90%) to concentrations of1 × 108 bacteria/ml, but adults were not affectedby cultures with 1 × 109 bacterial/ml. Traore et al.(unpubl. data) screened a wide range of isolates of

Bacillus thurgingensis, but found none effective inkilling banana weevil larvae.

f. Field trials and delivery systemsIn Guadeloupe, Delattre & Jean-Bart (1978) sprayedspores of B. bassiana at the base of banana mats inconcentrations of 2.2 × 1010 spores/ml in a plot ofPoyo (AAA) and 5 × 105–1 × 1011 spores/ml in a plotof Yangambi-Km5. None of the applications had anyeffect on banana weevil density.

In a field trial in Brazil, Mesquita (1988) appliedB. bassiana to pseudostem traps (120/ha) by immers-ing them in spore solutions with concentrations rangingfrom 8×107 to 6.48×108 spores/ml. Twenty-four appli-cations were made at 2-week intervals, with weevilscollected 15 days later and assessed for infection. A lowinfection rate, averaging 5%, was attributed to reducedspore viability under field conditions.

Batista Filho et al. (1991) screened five strainsof B. bassiana and made three applications of themost virulent in rice paste (50 ml per trap with 1 ×109 spores/ml) to pseudostem traps in a 1-ha stand andcompared trap catches to an adjacent 1-ha control. Theyfound 61% fewer adults and 91% fewer larvae in treatedtraps. However, in two other trials, Batista Filho et al.(1995b, 1996) found <20% control following periodicapplications of oil-based formulations of B. bassianaspores/ml to pseudostem traps.

In Costa Rica, Contreras (1996) studied the effectsof B. bassiana under field conditions with 600 m2 plots(126 plants). In half of the plots, B. bassiana wasapplied in two formulations (oil emulsion and rice sub-strate) to disk on stump traps, while in the other plots theentomopathogen was applied to pseudostem traps. Thedisk on stump traps captured 3–4 times as many weevilsas plots with pseudostem traps and, therefore, may bemore appropriate for disseminating entomopathogens.Immediately following application, the proportion ofinfected weevils was highest where oil formulationswere applied (72%) than for the rice-based formula-tion (48%) or controls (7%). In contrast, 8 days afterapplication, mortality was higher for the rice-based for-mulation (55%) than the oil formulation (28%) (thefigures we present here are estimated from graphs).

In Colombia, Castrillon (2000) applied B. bassianaand M. anisopliae to disk on stump traps in both rice-based substrates (15 g/trap) and in water suspension(15 cc/trap) through aspersion at each of 18 sites (onehigh and one low elevation site in each of nine munci-palities). Other treatments included a control, a ricesubstrate control (i.e. without entomopathogens) and


lorsban. Traps were evaluated weekly for 8 weeks,with weevils transported to the laboratory for isola-tion of pathogens. Application of entomopathogensin rice-based substrates resulted in 16% (site meanrange 4–34%) infection for B. bassiana and 16% (range0–42%) for M. anisopliae (i.e. sites weighted equally).In contrast, applications by aspersion resulted in only6% and 5% infection for B. bassiana and M. anisopliae,respectively. Surprisingly, up to 6% B. bassiana infec-tion was found in treatments where this pathogen hadnot been applied, while infection of M. anisopliae inother treatments was negligible. These data suggestthe either possibility of greater movement of weevilsinfected with B. bassiana than those infected withM. anisopliae or the existence of B. bassiana-infectedweevils in the population prior to evaluation.

In Ghana, Godonou et al. (2000) conducted fieldstudies to field efficacy and the spread of theB. bassiana following release of laboratory-infectedweevils. In the first experiment, 20 weevils werereleased at the base of recently planted plantain suckersthat had been: (1) protected by application of 60 g of oilpalm kernel cake containing 109 conidia/g; (2) coatedwith conidial powder containing 6 × 1010 conidia priorto planting; (3) planted without fungal application(controls). After 28 days, the suckers were uprootedand the number of weevils counted. A weevil recoveryrate of 23% with 30% infection was found on suckerscoated with conidial powder, compared to 31% recov-ery and 24% infection on suckers protected by oil palmkernel cake substrates and 50% recovery and no infec-tion on controls. Godonou et al. (2000) estimated 76%mortality in the two B. bassiana treatments, comparedto 1% in controls.

In a second experiment, Godonou et al. (2000)applied the same treatments to suckers planted amongmature plantain plants in an established field. Weevilswere trapped at the base of the suckers for 2 months,after which they were uprooted. Suckers protected withB. bassiana in oil palm kernel cake substrates hadthe highest mortality of trapped weevils (42%), low-est percentage of attacked plants (6%), fewest numberof larvae (6), and no dead suckers. In contrast, suck-ers coated with conidial powder had 6% dead weevilsin traps, 25% attack of plants, 17% plant death, and26 larvae. Controls displayed similar levels of attack asthose treated with conidial powder. From these results,Godonou et al. (2000) concluded that suckers couldbe protected at the critical stages of plant establish-ment by applications of conidial powder on an oilpalm kernel cake substrate. Under field conditions,

the fungus increased until the substrate was exhausted,providing extended protection.

In Uganda, Nankinga (1999) first tested traps in potsas a potential delivery systems for B. bassiana usingmaize culture, oil suspension and water suspension.Weevils were released, recaptured after 5 days andmaintained for 21 days. Maize culture produced a mor-tality of 83% in 21 days, compared to 47% in water sus-pension. The oil formulation killed 100% of the weevilsalthough only 60% showed signs of infection.

The same treatments were then applied to disk onstump and pseudostem traps under field conditions.After 5 days, weevils were collected from traps andmaintained in the laboratory for 21 days. Infectionrates were 50–60% for traps treated with maize cul-ture, 55–61% for traps treated with oil suspension,23–44% for traps treated with water suspension, and0% in oil and water controls. Moreover, captures werelower in traps treated with maize culture and oil sus-pension than for those treated with water suspensionor controls. Pathogenicity decreased in treated soilsafter 2 weeks, although 15% of weevils collected fromtreated soils showed signs of infection 5 months aftertreatment.

Nankinga (1999) applied 500 g of maize culture(2.65 × 108 spores/g) to the topsoil around bananamats in small (i.e. eight mats) plots covered with grassmulch. Weevil levels were then compared to plots inwhich no fungi were applied. Four weeks after appli-cation, 48% of collected weevils in treated plots wereinfected. Moreover, 20% of weevils collected in treatedplots 5 months after treatment were infected. However,the treatment did not reduce weevil trap captures overa 7-month period suggesting migration across smallplots.

In a second experiment, weevil populations weremonitored with pseudostem traps for 8 months inplots that received two B. bassiana applications. Meanweevil counts were lowest in plots treated with maizeformulation (40 trapped weevils per plot) followed byplots receiving soil formulation (54), oil formulation(68), and controls (81). The incidence of field mortalityof weevils observed in traps was low with a maximumof 5% and often under 1%. Peak mortality reached 15%in plots treated with maize formulation and 13% in soilformulation. Maize-based formulations also tended toreduce weevil damage levels in the central cylinderand cortex. Although the maize formulation showedthe potential of field level control, the application rate(250–500 kg/ha; 1×1014–2×1015 spores/ha) employedin this study was not economically viable.


Sublethal effects of entomopathogens against insectadults may lead to reduced fecundity or egg sterility(Nankinga 1999). For example, Nankinga (1999) foundthat application of B. bassiana to the soil in pots led toa 56% infection of adults and a 73% reduction in eggnumber. In another experiment, where suckers weretreated with B. bassiana, egg eclosion was 39% lessthan in controls. In addition, infected adults can trans-mit both B. bassiana and M. anisopliae to eggs (andsubsequent larvae) (Nankinga & Ogenga-Latigo 1996).

Disk on stump and pseudostem traps may aggre-gate weevils at delivery sites for entomopathogens(Kaaya et al. 1993; Contreras 1996; Nankinga 1999).Budenberg et al. (1993a) further suggested that semio-chemicals might increase weevil attractiveness ofentomopathogen-baited traps. This would require amodification of the current pheromone-based pitfalltrap design such that the weevils become infectedrather than drown. Such a method would be advan-tageous over standard pitfall trapping only if infectedadults were able to transmit the pathogen to otherweevils. Currently, IITA, Uganda NARO, and ICIPEare also trying to develop a delivery system forentomopathogens using a kairomone-based trappingsystem.

g. Transmission among weevilsNankinga (1994) exposed adult weevils to B. bassiana-infected weevils. Transmission rates were higher fromdead weevils than from live infected weevils. For exam-ple, a single dead weevil could infect 70% of exposedweevils, while a living weevil infected 28%. If unin-fected weevils were exposed to three dead or threeinfected live weevils, infection rates increased to 98%and 70%, respectively. Schoeman & Schoeman (1999)mixed uninfected weevils with weevils inoculated withB. bassiana spore-suspensions. After 44 days, all ofthe exposed weevils were dead and showed signs ofmycosis, while 24% of the uninoculated weevils werealso dead with signs of mycosis. However, transmissionrates among weevils in field situations, where weevildensity is relatively low, remains unclear.

2. Endophytes

A wide variety of endophytic fungi have been isolatedfrom nearly all examined plants (ranging from grassesto trees) and plant tissues (Carroll 1991). Many of thesehave developed mutualistic relationships with plants

and some act as antagonists to pests and diseases. Endo-phytes may be classified as constitutive or induciblemutualists; the former occur throughout the life of theirhosts, while the latter remain in a latent state untilstimulated by pest attack (Carroll 1991). The preva-lent mode of action appears to be through the pro-duction of metabolites that act as oviposition repel-lents, toxins or feeding deterrents. Plant physiologicaland ecological factors may influence endophyte effi-cacy. Endophytes can enhance resistance to specialistherbivores that have evolved mechanisms to circum-vent the plant’s normal defences (Carroll 1991; Breen1994).

Research is currently being undertaken at IITAin Uganda, in collaboration with the University ofBonn, for the development of a biological controlprogramme using endophytes against banana weevil.Research protocols include: (1) isolation and identi-fication of endophytes from banana corms and pseu-dostems; (2) screening against banana weevil eggsand larvae; (3) determination of mechanisms by whichpromising strains kill weevil immatures; (4) reinocula-tion into banana tissue culture plantlets and/or bananasuckers; (5) determination of distribution, prevalence,and persistence of promising endophytes within bananaplants; (6) efficacy studies in pot and field trials for arange of clones and under different ecological condi-tions; (7) developing markers or vegetative compati-bility groups (VCG) to identify strains of endophytes;(8) pathogenicity testing of promising strains in bananaand other crops.

Griesbach (1999) obtained 200 isolates from atotal of recently harvested 64 plants on 21 farmsin Ntungamo district, Uganda. Samples were takenfrom five highland cooking bananas (AAA-EA), onehighland brewing banana (AAA-EA) and the exoticclone Kayinja (ABB). Spore suspensions of 12 iso-lates (8 Fusarium spp., three Acremonium spp., oneGeotrichium sp.) caused 80–100% mortality in weevileggs, while 74 additional isolates caused 60–79% mor-tality. Further work was restricted to the 12 mostpromising isolates. Testing of mycotoxins (ratherthan direct colonisation) produced 30–88% mor-tality for the Fusarium isolates and 16–24% forAcremonium. Screening against banana weevil larvaegave 0–48% mortality with the best two strains beingF. cf concentricum (48%) and F. oxysporum (32%).It is possible that endophytes might induce resistancein banana plants to pests, although initial studies oninduced effects against banana nematodes producednegative results (Niere 2001).


Griesbach (1999) was able to successfully inocu-late tissue culture plants with endophytes. Colonisationrates were 39–73% colonisation for Fusarium,0–26% colonisation for Acromonimum, and 0% forGeotrichium. For the best Fusarium strains, inoculationsuccess was 38% for Valery (AAA), 44% for Kayinja,73% for Nabusa (AAA-EA), and 88% for Gros Michel(AAA). Within three highland cooking clones, the bestFusarium strains, colonisation rates were 12% in roottips, 39% in root bases, 48% in corms, and 3% in pseu-dostems. By contrast, there was only 3% establish-ment of inoculated endophytes in pared or hot watertreated suckers. Preliminary pot trials on the effectsof inoculated endophytes on weevil damage in tis-sue culture plants produced promising but inconsistentresults.

3. Entomopathogenic nematodes

The use of entomopathogenic nematodes for insectcontrol and, specifically, against banana weevil hasbeen reviewed by Treverrow et al. (1991), Parnitzki(1992), and Schmitt (1993). The most commonlyused species are within the genera Steinernema andHeterorhabditis. These have received wide attention asbiological control agents because of wide host range,ability to kill host rapidly, and no adverse effects onenvironment (Schmitt 1993).

The infective stage locates its host by detect-ing excretory products, temperature gradients, etc.(Schmitt 1993). Five species of Xenorhabdus bacteriaare mutualistically associated with Steinernema whilePhotorhabdus spp. is associated with Heterorhabditis.Infective juvenile nematodes enter through natu-ral orifices (Steinernema) or interskeletal membrane(Heterorhabditis) (Treverrow et al. 1991). After enter-ing the host, the nematodes penetrate mechanically intothe haemocoel and release Xenorhabdus which causessepticaemia and insect death within 1–2 days (Schmitt1993).

Entomopathogenic nematodes have a non-feedingstage that can survive in the soil for extended peri-ods. Soil temperature, soil moisture, and soil types areimportant abiotic factors which affect these nematodesurvival and performance (Schmitt 1993). Parnitzki(1992) suggested that sensitivity to drought, high tem-peratures, and ultraviolet light are also limiting factorsin the efficacy of entomopathogenic nematodes.

In surveys in Brazil, Schmitt (1993) found ento-mopathogenic nematodes to be widespread. However,naturally occurring mortality of banana weevils to

entomopathogenic nematodes was low and, as withentomopathogenic fungi, viable delivery systems arean important consideration.

Entomopathogenic nematodes (Steinernema spp.and Heterorhabditis spp.) have been tested againstbanana weevils in Australia and the Pacific (Treverrowet al. 1991; Parnitzki 1992; Treverrow & Bedding1991; Treverrow 1993, 1994), the Caribbean (Laumondet al. 1979; Kermarrec & Mauleon 1975, 1989;Figueroa 1990), Florida (Pena et al. 1993) and Brazil(Schmitt 1993). Entomopathogenic nematodes maybe more effective against banana weevil larvae thanagainst weevil adults (Figueroa 1990; Kermarrecet al. 1993; Pena & Duncan 1991; Treverrow 1994).For example, Figueroa (1990) reported Steinernemafeltiae, S. glaseri, and S. bibionis caused 13–66%mortality of late-instar larvae in laboratory assaysand 100% mortality and a 70% reduction in weevilgalleries in potted plants. Pena et al. (1993) found47–89% mortality of weevil larvae compared to 45%on weevil adults. Larval mortality in greenhousetests was 37%. Treverrow et al. (1991) found ento-mopathogenic nematodes applied to crosscuts in resid-ual corms were equally effective against small and largelarvae.

However, the cryptic habitat of weevil larvae withinliving plants makes delivery against these stages diffi-cult (Treverrow 1994). For example, Treverrow et al.(1991) and Treverrow & Bedding (1993) found spray-ing of entomopathogenic nematodes onto corms inef-fective against larvae as there were few entry pointsand the holes made by adults were quickly blocked bycallus tissue. In addition, it is difficult to know whichplants are infected with larvae, such that applicationscan not be restricted to plants or plots with high weevilnumbers.

Therefore, Parnitzki (1992) and Treverrow (1993,1994) recommended that applications of ento-mopathogenic nematodes should target adult weevils.Parnitzki (1992) screened 30 strains of Steinernemaand Heterorhabditis against banana weevil adults inTonga and Australia and found that the most effectivestrains differed between the two sites. This suggestseither strain–environment interactions or the presenceof weevil biotypes. Parnitzki (1992) recommended thatstrains should be screened locally before implementinga programme using entomopathogenic nematodes. InTonga, Parnitzki (1992) felt that only one Steinernemaand three Heterorhabditis offered potential. Severalother strains were able to locate the weevil but wereunable to overcome the host’s defences.


The development of a delivery system had to con-sider a number of factors. First, infection of adultweevils was limited by difficulties the nematodes expe-rienced in entering the host (Treverrow & Bedding1993). In laboratory tests, the first spiracle appearedto offer the best entry point, but this site is secluded bythe insect’s elytra. Second, infection required at leastseveral days of exposure to the nematodes, with max-imum efficacy attained with 7–14 days exposure. Thismeant that an ideal delivery system would retain weeviladults at the exposure site.

Ground sprays were considered in that they arenot limited by the availability of certain plant stagesor residues, but these required high nematode den-sities and persistence was limited (Treverrow 1994).In contrast, the efficacy of applications could also beincreased if adults could be aggregated and retainedat delivery sites (i.e. by attraction to traps). In Brazil,for example, Schmitt et al. (1992) baited pseudostemand disk on stump with S. feltiae and compared theseto ground applications. At 7, 14, and 21 days, mor-tality was higher on pseudostem traps (51, 40, 40%,respectively) and disk on stump traps (70, 51, 32%)than following soil applications (58, 24, 25%).

In Tonga, Parnitzki (1992) applied entomopathogenicnematodes to cuts in corm stumps (left at 0.5 m) shortlyafter harvest. Although weevil adults were attracted tosuch sites, mortality was low (i.e. 20%) and there wasno reduction in damage levels.

Treverrow et al. (1992) and Treverrow & Bedding(1993) developed a delivery system for ento-mopathogenic nematodes capitalizing on the weevil’sattraction to cut corms and damaged plants. Initially,the made two holes in the residual corm or split it par-allel to the ground (Treverrow 1993). This was laterrevised to the use of two conical shaped cuts in residualcorms. These cuts attracted adult weevils and providedthigmotactic stimuli that encouraged them to remain atthe infection sites. The holes also buffered the deliverysite against temperature extremes and provided excel-lent conditions (high humidity, moderate temperatures,protection against ultraviolet light) for nematode per-sistence. The nematodes were released at a density of250,000 per hole in a formulation including a poly-acrylic gel (to reduce water build-up and incidence ofnematode drowning) with an adjuvant of 1% paraffin oil(to encourage the weevils to raise their elytra, expos-ing the first spiracle for nematode entry). The nema-todes persisted for up to 50 days and attacked bothadults and larvae (Treverrow et al. 1991; Treverrow1994). At moderate weevil infestation levels, nematode

baits performed as well or better than insecticides(Treverrow 1993; Treverrow & Bedding 1993), butwere not as effective as pesticides in heavily infestedfields (Treverrow 1994). However, controls based onentomopathogenic nematodes were not economicallycompetitive with pesticides (Treverrow 1993, 1994).

In contrast to the positive results obtained byTreverrow, Smith (1995) reported injection of ento-mopathogenic nematodes into cuts from the pseu-dostem to the corm in mature plants and residues gaveno benefit over the control. Whereas he was unable toapply a gel, he attributed the lack of control to larvaldrowning. However, in later trials in which the gel wasadded to the application formulation, he again foundno benefit. Moreover, he noted that the system was notattractive to farmers.

XIV. Host Plant Resistance

The literature on the susceptibility of Musa clonesto banana weevil attack is largely fragmentary withhighly variable and often contradictory findings(Pavis & Lemaire 1997; Kiggundu et al. 1999).Most often, reported results reflect comparisonsamong a small number of clones used in field tri-als (Sen & Prasad 1953; Hord & Flippin 1956;Moreira 1971; Oliveira et al. 1976; Mitchell 1978;Zem et al. 1978; Haddad et al. 1979; Viswanath1981; Ittyeipe 1986; Irizarry et al. 1988; Kehe 1988;Bakyalire 1992; Batista Filho et al. 1992; Minost 1992;Pavis 1993; Seshu Reddy & Lubega 1993; Speijeret al. 1993; Davide 1994; Pone 1994; Stanton 1994;Vittayaruk et al. 1994; Abera 1997; Mestre & Rhino1997; Silva & Fancelli 1998).

Fogain & Price (1994), Ortiz et al. (1995), Rajamonyet al. (1993, 1994, 1995), Anitha et al. (1996) andKiggundu (2000) conducted screening trials to identifyexisting clones displaying resistance to banana weevil.These results have been reviewed by Pavis & Lemaire(1997), Kiggundu et al. (1999), and Kiggundu (2000).

The variability in susceptibility reported by differ-ent authors for closely related clones, or even acrossgenome groups may reflect differences in samplingmethods for assessing weevil damage. In field sur-veys, for example, Gold et al. (1994a) and Bosch et al.(1996) found different trends when comparing damageto the corm surface, the outer cortex and the centralcylinder. In Ugandan surveys, plantains (AAB) andhighland bananas (AAA-EA) appeared more suscepti-ble to banana weevil attack than other genome groups


(Gold et al. 1994a). For example, both plantains andhighland bananas displayed high levels of attack on thecorm surface with considerable penetration into cortexand central cylinder. In contrast, weevil attack of GrosMichel (AAA) was restricted to the corm surface andcortex with limited penetration into the central cylin-der. Other introduced beer (AB, ABB), cooking (ABB),and dessert clones (AB) were relatively resistant withlittle surface damage and virtually no penetration intothe corm.

In the Kagera region of Tanzania where bananaweevil damage is often severe, Bosch et al. (1996)found high levels of surface corm damage in endemichighland cooking bananas (AAA-EA) as well as exoticAAA, AB, and ABB clones. However, only the high-land group sustained high levels of weevil penetra-tion into the cortex and cylinder. Ogenga-Latigo &Bakyalire (1993) found that Ndiizi (AB) had similarlevels of surface damage to that of highland cookingbanana, but only 16% as much internal damage.

Unfortunately, some researchers have assessedweevil damage to the corm surface, while others haveestimated damage to the interior of the corm, makingresults difficult to interpret and compare. For exam-ple, in a screening trial in Uganda, Kiggundu (2000)found that Nsowe (AAA-EA) scored highest amonghighland banana clones in damage to the corm surfacebut lowest in internal corm damage. Pavis & Lemaire(1997) and Mestre (1997) noted the need for standardscreening methods and reference cultivars. Kiggundu(2000) recommended the use of total cross section dam-age (c.f. Gold et al. 1994a) as this measure had a highlevel of heritability and was well correlated with otherindices of weevil damage. In contrast, Rukazambugaet al. (1998) suggested the use of damage to the centralcylinder as this damage appeared to have the greatestimpact on plant growth and yield.

Variable findings from studies conducted in dif-ferent locations may also reflect ecological differ-ences or genetic variability (i.e. biotypes) amongweevils (Pavis & Lemaire 1997; Kiggundu et al. 1999).However, some general trends do appear.

1. Resistance across genome groups

Of the two wild progenitors of edible bananas,M. acuminata (AA) has been reported as more suscepti-ble to banana weevils than M. balbisiana (BB) (Saraiva1964; Simmonds 1966; Vilardebo 1973; Mesquita et al.1984). However, both of these diploids escaped attackin a screening trial in Cameroon (Fogain & Price 1994).

Mesquita et al. (1984) further suggested that geneticcontribution from M. balbisiana in naturally derivedor bred hybrids conferred higher levels of resistance toweevils.

Nevertheless, plantains (AAB) are generally con-sidered the most susceptible Musa genome group tobanana weevil attack and much more susceptible thanmost AAA dessert bananas (Ghesquiere 1925; Pinto1928; Simmonds 1966; Haddad et al. 1979; Mesquitaet al. 1984; Ittyeipe 1986; Jones 1986; Pavis 1988;Bakyalire 1992; Seshu Reddy & Lubega 1993; Speijeret al. 1993; Fogain & Price 1994; Gold et al. 1994a;Price 1994; Sponagel et al. 1995; Pavis & Lemaire1997). In India, however, Viswanath (1981) reportedplantains as resistant to banana weevil.

Highland cooking bananas (AAA-EA) are also con-sidered highly susceptible to banana weevil (Sikoraet al. 1989; Bakyalire 1992; Gold et al. 1994a; Boschet al. 1996; Rukazambuga et al. 1998). Reports on sus-ceptibility of AAA dessert bananas (e.g. Gros Michel,Cavendish, Williams, Valery) have ranged from resis-tant to susceptible (Zem et al. 1978; Viswanath 1981;Mesquita et al. 1984; Mesquita & Caldas 1986;Fogain & Price 1994; Gold et al. 1994a; Stanton1994; Sponagel et al. 1995; Bosch et al. 1996). Ostmark(pers. comm.) and Sponagel et al. (1995) suggest that inAAA dessert bananas weevils favour crop residues overdeveloping plants and are thus unimportant. In contrast,Viljoen (pers. comm.) reported serious banana weevilproblems on Cavendish bananas on the southeasterncoast of South Africa.

Ortiz et al. (1995) reported that wild diploids weregenerally more resistant than polyploids. AB and ABBbananas are often considered among the most resis-tant Musa clones to banana weevil (Hord & Flippin1956; Mesquita et al. 1984; Mesquita & Caldas 1986;Seshu Reddy & Lubega 1993; Gold et al. 1994a;Musabyimana 1995; Ortiz et al. 1995; Bosch et al.1996; Abera 1997). Haddad et al. (1979) foundABB clones to be intermediate in susceptibility tobanana weevil between plantains and dessert bananas,while Viswanath (1981) found ABB bananas to besusceptible.

Limited information is available on susceptibility oftetraploids to banana weevil. Ittyeipe (1986) reportedAAAA clones to be the most susceptible to weevilattack, while Viswanath (1981) found larval successgreatest on AABB bananas.

In germplasm collections in Cameroon and Nigeria,ensete appeared to be highly susceptible to bananaweevil (Pavis & Lemaire 1997; C. Gold pers. observ.).


However, in Ethiopia, where the crop is most widelygrown, ensete largely escapes attack because mostproduction is above the weevil’s upper elevationalthreshold (M. Bogale et al. unpubl. data).

Kiggundu (2000) conducted a screening trial inUganda with 45 clones including representatives fromall five clonal groups of East African highland bananas(c.f. Karamura 1998), plantains, exotic cooking, andbrewing (ABB), dessert (AAA), diploids (AA, AB)and hybrids. Cross section damage (c.f. Gold et al.1994b) ranged from 0% to 11%. Plantains and highlandbananas appeared most susceptible followed by ABBs,hybrids, ABs, AAAs, and AAs (Table 9). Cluster anal-ysis suggested that 19 clones were highly suscepti-ble to weevil attack (mean damage 8%), 17 cloneswere intermediate in susceptibility (4%) and 9 cloneswere resistant (1%) (Table 10a).

In India, Rajamony et al. (1993, 1994, 1995) andAnitha et al. (1996) screened 87 clones (including7 AA, 7 AB, 18 AAA, 27 AAB, and 28 ABB) col-lected from 13 localities. Weevil damage was ranked0–4 (although the scoring method was not clearlydescribed). Considerable variability was found withineach genome group with lowest damage in the ABbananas and limited differences among the othergroups. In this study, there was little relationshipbetween susceptibility to banana weevil and to banananematodes.

In summary, the data suggest that plantains (AAB)and highland cooking banana (AAA-EA) are most sus-ceptible to banana weevil attack. Diploids, other AAA(e.g. dessert) bananas and AB and ABB bananas appear

Table 9. Means (±standard error) banana weevil damage bygenome groups of Musa in Uganda

Genome Musa type Mean total Range ofgroup weevil total weevil

damage damage

AAB Plantains 7.8 7.5–8.1AAA-EA East African 5.9 2.7–9.9

highlandbananas

ABB Kayinja 3.3 2.3–4.1Bluggoe

Hybrids Plantain derived 6.6 6.3–7.9Banana derived 0.2 0.1–0.2

AB Ndiizi, Kisubi 2.4 1.0–3.1AAA Yangambi-km5, 1.8 0.4–4.0

Cavendish,Gross Michel

AA Wild banana 0.2 0.2Calcutta-4

Source: Kiggundu (2000).

to be less susceptible, although considerable variabilityhas been reported from different studies.

2. Clonal resistance

In screening trials in Cameroon (Fogain & Price 1994)and Nigeria (Ortiz et al. 1995), all plantain clonesappeared susceptible to banana weevil. In contrast,Chavarria-Carvajal (1998) evaluated 8 plantain clonesand found the Common dwarf variety and a Lacknauclone to have less than 20% of the damage occur-ring in Sin Florescencia and Rhino Horn plantains.Irizarry et al. (1988) and Fogain & Price (1994)also found Lacknau clones less susceptible than otherplantains.

In Uganda, field survey data suggested differencesin susceptibility to banana weevil attack among high-land banana clones (Gold et al. 1994a). Atwalira(=Nassaba) and Kisansa displayed weevil damagescores 2–3 times higher than those for Mbwazirumeand Nakyetengu, while the degree of penetrationinto the central cylinder was greatest for Nakitembe,Namwezi, and Musakala. However, these resultswere biased by clonal distribution. For example,Mbwazirume was quite common on farms in regionswith high levels of management and commercial objec-tives (e.g. Masaka, Mbarara districts). In contrast,Atwalira was often grown on small farms with lim-ited inputs (e.g. Luwero). Moreover, Atwalira primarilyoccurred in 3 sites, all of which supported high levels ofweevil damage. Further analysis, employing Z values(i.e. standard scores) (Zar 1984) to eliminate site dif-ferences, showed Atwalira to have only slightly aboveaverage levels of damage in the sites where it occurred(Gold et al. unpubl. data).

Kiggundu’s (2000) screening trial included 26highland bananas among the 45 evaluated clones.Damage scores within the highland banana groupranged from 3% to 10%. Cluster analysis of allclones suggested that 15 highland clones were sus-ceptible, while 11 were intermediate in susceptibil-ity (Table 10a). However, analysis of only the high-land group suggested that 7 clones were highly sus-ceptible to weevil attack (mean damage 9%), 13clones were intermediate in susceptibility (6%) and6 clones were resistant (4%) (Table 10b). Of themost popular and widespread clones, Mbwazirume andNakyetengu appeared relatively resistant. One brew-ing clone was considered susceptible, three were inter-mediate in susceptibility and two appeared relativelyresistant.


Table 10a. Three banana-weevil susceptibility response groups derived from cluster analysisof 45 Musa cultivars in a screening trial in IITA Sendusu Farm, Namulonge, Uganda

Resistant (Cluster 1) Intermediate (Cluster 2) Susceptible (Cluster 3)

Name Total Name Total Name Totaldamage damage damage

TMPx15108-6 2.0 Nakamali 6.4 TMPx7152-2 10.7Cavendish 1.7 Enshenyi 5.5 Kibuzi 10.1TMBx612-74 1.4 Kabula 5.4 Ndiibwabalangira 9.9Kisubi 1.0 Siira 5.2 Endiirira 9.2Yangambi-Km5 0.3 Nandigobe 5.2 Nakawere 8.8TMB2x8075-7 0.3 Mutangendo 4.9 Obino l’Ewai 8.3Calcutta-4 0.2 Bukumu 4.9 TMPx5511-2 7.9TMB2x7197-2 0.1 Nakyetengu 4.1 Namafura 7.7TMB2x6142-1 0.1 Bluggoe 4.0 Atwalira 7.7

Bogoya 3.7 Namwezi 7.6Nsowe 3.3 Naminwe 7.6Mbwazirume 3.1 Gonja 7.3Nalukira 3.1 TMPx7002-1 6.8Tereza 3.0 Musakala 6.5Ndiizi 2.9 Nakabululu 6.4Kayinja 2.4 Shombobureku 6.4FHIA03 2.2 Nakitembe 6.2

Bagandeseza 6.1Kisansa 5.8


Table 10b. Three response groups derived from cluster analysisof EAHB cultivars in a screening trial in IITA Sendusu Farm,Namulonge, Uganda (∗ = brewing types)

Resistant Intermediate Susceptible

Mbwazirume Bagandesesa∗ AtwaliraNakyetengu Enshenyi Endiirira∗

Mutangendo Kabula∗ KibuziNsowe∗ Kisansa NaminweNalukira∗ Bukumu NakawereTereza Musakala Ndiibwabalangira

Nakabulu NamafuraNakamaliNakitembeNamweziNandigobeShombobureku∗

Siira


3. Mechanisms conferring resistance

Successful attack of bananas by banana weevilsinvolves host plant location, host plant acceptance(oviposition), and host plant suitability (larval survival,developmental rate, and fitness). Host plant resis-tance may affect any of these processes. Most com-monly host plant resistance mechanisms have beenattributed to antixenosis (non-preference), antibiosisand/or host plant tolerance (Painter 1951). For banana

weevil, available data suggest that antibiosis is themost important factor conferring host plant resistance,while antixenosis is of little importance. Little has beenreported on host plant tolerance to banana weevil attackas such work would require yield loss studies overseveral crop cycles.

a. AntixenosisAntixenosis suggests that resistant clones avoid pestattack by reducing rates of host plant location(i.e. attraction) and/or host plant acceptance; the com-bined effects of these two processes would be reducedoviposition. Pavis & Lemaire (1997) suggested thatantixenotic factors may also deter adult feeding.

Rwekika (1996) and Rwekika et al. (2002) foundthe phenolic glucoside salicin an attractant to bananaweevils and suggested that it served as a feeding stim-ulant for adults. He further noted that salacin andglucose were present in higher levels in susceptiblehighland banana (AAA-EA) clones than in resistantclones such as Ndiizi (AB), Pisang awak (ABB), andKivuvu (ABB). Gowen (1995) reported all clones tobe susceptible to banana weevil attack and suggestedthat differences in damage levels reflected differencesin weevil attraction.

The data on host plant attraction to susceptible andresistant clones is equivocal. In laboratory choice tests,


Mesquita et al. (1984) found clonal preferences foradult feeding which differed from those for oviposition,suggesting different levels of host plant acceptance.Minost (1992) found that Burmanica (AA) was mostattractive to adult weevils followed by Pisang awak(ABB), Borneo (AA) and French Clair (AAB), whilePetit Naine (AAA), and Rose (AA) were much lessattractive. However, clonal attraction was not related toweevil damage: Burmanica had low levels of periph-eral damage, Petit Naine had intermediate damage,while Pisang awak and French Clair had high dam-age. Sumani (1997) also found attraction to pseu-dostems and corms in choice chambers did not reflecthost plant susceptibility. Minost (1992) concluded thatresistance mechanisms must be related to oviposi-tion and larval development rather than to host plantattraction.

In contrast, Budenberg et al. (1993b) reported thatfemales were equally attracted to cut corms andvolatiles from resistant and susceptible cultivars. Hesuggested that host plant attraction was related to adultfeeding and not selection of oviposition sites. However,adults were more commonly observed feeding on rot-ting banana tissue (e.g. residues, decaying leaves).Pavis & Minost (1993) also found similar levels ofattractivity to resistant and susceptible clones.

In field studies, Musabyimana (1995) observed dif-ferential attraction (based on trap capture rates at thebase of the mat) among clones; However, trap captureswere not related to subsequent damage. In a screeningtrial in Uganda, Kiggundu (2000) found some differ-ence in trap captures among clones, but that many of theresistant exotic clones and hybrids had high numbersof weevils, while some of the more susceptible high-land cooking clones (e.g. Atwalira) had low capturerates. As a result, there was no relationship between trapcatches and subsequent damage. Abera (1997) foundsimilar trap captures at the base of mats of the resistantclone Pisang awak (ABB) as for 5 susceptible highlandbanana (AAA-EA) clones.

Little work has been done on host plant accep-tance. Abera (1997) and Abera et al. (1999) foundfield oviposition on Kayinja (ABB) to be similar to thaton highland banana clones, even though the latter dis-played much higher levels of weevil damage. Kiggundu(2000) looked at oviposition on resistant and suscepti-ble clones in both choice and no-choice experiments.There was very little mean separation and the lowerlevels of oviposition occurred on clones (Atwalira,Nakyetengu, Muvubu) that were not consideredresistant.

In summary, the banana weevil is a relatively seden-tary insect living in perennial systems with an abun-dance of host plants. It is unclear to what extent weevilsare preferentially attracted to one clone over another.Data on movement patterns suggest that some weevilsmay spend extended periods of time at the base of a sin-gle mat, while less than 40% of the weevils moved morethan 10 m in 7 weeks (Gold et al. 1999d). Kiggundu(2000) suggested that it is unrealistic to think thatbanana weevils might walk far looking for a suitablehost. Most likely, tenure time at the base of any givenmat is more related to environmental factors such assoil moisture.

Although some authors suggest that resistant cloneshave feeding deterrents (Pavis & Lemaire 1997) or lim-ited quantities of feeding stimulants (Rwekika 1996),there is little evidence to suggest that adult feeding isan immediate prerequisite for oviposition. The weevilcan live for extended periods of time without feedingand can move freely from preferred feeding sources(e.g. decaying residues) to oviposition sites. Availabledata indicate that the weevil will freely oviposit onboth susceptible and resistant clones, suggesting thatantibiosis plays a more important role in host plantresistance (Abera 1997; Kiggundu 2000).

b. AntibiosisAntibiotic factors are those which negatively influencelarval performance (i.e. poorer survivourship, slowerdevelopment rates, reduced fitness). These factors mayinclude physical (e.g. sticky sap and latex, corm hard-ness), antifeedants, toxic secondary plant substancesand nutritional deficiencies.

In a mixed cultivar trial, Abera (1997) found weevildamage to the interior of the corm to be 5–25 timeshigher in 5 highland banana clones (3 cooking and2 brewing) than in Pisang awak. Banana weevil attrac-tion and oviposition on Kayinja was similar to that onthe highland bananas, while larval survivourship wasestimated as 10–23 times higher in highland bananasthan in Kayinja. From these data, Abera (1997) con-cluded that antibiosis explained why Kayinja was rel-atively resistant to banana weevil.

Mesquita & Alves (1983), Mesquita et al. (1984),and Mesquita & Caldas (1986) found that bananaweevil immatures developed faster and had fewerecdyses on some clones than on others. In three studies,the larval period ranged 22–29, 25–32, and 35–44 days,depending on clone. Banana clone also influencedthe duration of the pupal stage and pupal weights insome but not all of the trials. Lemaire (1996) reported


slower larval development and higher larval mortalityon the resistant clone Yangambi-Km5. Silva & Fancelli(1998) also reported the influence of clone on larvaldevelopmental period.

Kiggundu (2000) found the length of the larvalperiod ranged 29–40 days for weevils reared on differ-ent clones. Two resistant clones, FHIA-03 and Kayinja,increased larval developmental time. Eclosion rateswere similar among clones. Larval mortality ranged 5–100% with highest levels occurring in resistant clonessuch as Kayinja (100%) and Kabula (AAA-EA) (90%).Larvae reared on Mbwazirume (AAA-EA), FHIA-03, Ndiizi (AB), Yangambi-Km5 (AAA) also hadhigh mortality rates. Within the East African highlandgroup, larval mortality on the more resistant brew-ing clones tended to be higher than on the cookingbananas. Corm extracts from Kayinja applied to sus-ceptible corm material had little impact on eclosion, butseverely inhibited larval feeding, while extracts fromother clones did not.

Pavis & Minost (1993) found a negative correla-tion (r = −0.47) between corm hardness and infes-tation rate and hypothesised mechanical resistanceto oviposition or larval development. Ortiz et al.(1995) assessed 5 plantains cultivars, 2 AAA dessert,Calcutta 4 (AA), Bluggoe (ABB), Fougamou (ABB)plus 97 euploid hybrids for corm hardness in transverseand longitudinal sections within 1 h after collection.All plantains were equally susceptible to the weevil,while significant differences were found among theeuploid hybrids for weevil levels and corm hardness;phenotypic correlations were not significant betweencorm hardness and weevil damage scores in segregat-ing progenies suggesting other resistance mechanismsmay be more important. Kiggundu (2000) also foundno phenotypic relationship between corm hardness andweevil damage.

Kiggundu (2000) screened extracts of 15 clones from3 weevil response levels, using high performance liquidchromatography. The chromatograms displayed peakson resistant AB or ABB clones not found on susceptibleclones or resistant AAA clones (i.e. Yangambi-km5 andCavendish). The data suggest that an antibiotic mecha-nism (e.g. toxic compound) is present in cultivars withthe B genome, while resistant AA or AAA bananasmay have other mechanisms of resistance.

Methanol extracts from fresh corm of two cultivars,Kayinja (resistant) and Atwalira (AAA-EA, suscepti-ble), were then bioassayed for their effect on weevillarvae. Application of the Kayinja extract to nutrientmedia resulted in a significantly lower developmental

rate and higher mortality of early-instar larvae, than thatresulting from the Atwalira extract (Kiggundu et al.unpubl. data). A bioassay-guided separation processof the Kayinja extract was then undertaken using chro-matographic techniques; 2 of the 16 fractions obtainedwere found to be very active against weevil larvae. Todate, however, the compounds responsible for activityagainst banana weevil have not been identified.

c. ToleranceTolerance suggests that the host plant can sustainhigh levels of insect damage without yield reduction.Cuille & Vilardebo (1963) suggested that Gros Michel(AAA) was resistant because the large size of the cormconferred tolerance to weevil attack. Kiggundu (2000)also suggested that corm can reduce the proportion ofdamaged tissue. Pavis (1993) suggested that the vigor-ous growth of Pisang awak allowed it to tolerate moder-ate levels of attack. However, no studies have compareddamage thresholds and related yield losses for differentMusa clones.

4. Breeding for resistance

To date, there have been no attempts to breed bananasor plantains for resistance to banana weevil. Breedingfor resistance depends upon sound knowledge of resis-tance mechanisms, resistance markers and the geneticsof resistance (Kiggundu et al. 1999). As a foundation, itis important to determine if there are useful sources ofresistance within the available germplasm. Kiggundu(2000) observed that the wild diploid Calcutta-4 andthe clones Yangambi-Km5 and FHIA-03 showed highlevels of resistance and might be exploited in breed-ing programmes. Lemaire (1996) and Mestre & Rhino(1997) also found Yangambi-Km5 to be highly resis-tant to banana weevil. Calcutta-4 has already been suc-cessfully used in conventional breeding programmesin Nigeria and Uganda, while the male/female fertil-ity of Yangambi-Km5 and FHIA-03 still need to bedetermined (Kiggundu 2000).

Alternatively, breeding for resistance to bananaweevil can employ the use of biotechnological toolsto identify and introduce resistant genes into plantainsor highland banana plants. Such studies could seek toidentify candidate genes from within and without theMusa genome and through currently available transfor-mation systems, study the expression of such genes inbanana. This type of research could also include both


damage-induced gene expression (i.e. in known resis-tant Musa cultivars) and enzyme (amylase and pro-tease) inhibitor gene expression using foreign genesof plant origin.

XV. Botanicals

Walangululu et al. (1993) reported that Tephrosialeaf powder reduced weevil attraction to treated baits.However, no follow up studies to these preliminaryresults have been reported. In contrast, McIntyre et al.(2002) found that intercropping and mulching withT. vogelii had no effect on either weevil adult popula-tion size or weevil damage to highland cooking banana.

In Kenya, Musabyimana (1999) and Musabyimanaet al. (2001) conducted a systematic and detailed studyon the effects of neem (Azadirachta indica) seed deriva-tives on banana weevil adult activity , success of imma-tures and resulting damage. This research, includingboth laboratory and field trials, employed different for-mulations of neem seed powder (NSP), neem kernelpowder (NKP), neem cake (NC), and neem oil (NO).These formulations were derived from ripe fruits fromcoastal Kenya which were then dried for 3–4 days to13% moisture content. The azadirachtin content wasdetermined as 4000 ppm for NSP, 5500 ppm for NKP,5800 ppm for NC, and 850 ppm for NO. Musabyiamana(1999) and Musabyimana et al. (2001) reported thefollowing results:

Adult settling: In laboratory choice and no-choiceexperiments, treatment of corm pieces(cv Nakyetengu, AAA-EA) with NSP and NO greatlyreduced the settling response of adults at 48–84 h afterapplication. The strength of this response was dosedependent. For example, in a choice experiment, 53%of the weevils settled on controls compared to 11% on2.5% NO formulation and 6% on 5% NO formulation.

Oviposition and hatchability: Oviposition was 3–10times greater on controls than on treated corm. Hatch-ability of inserted eggs in controls (41%) was 3–20times treater than on treated material.

Larval feeding: Third-instar larvae took longer tolocate feeding sites and initiate feeding on neem-treatedcorm disks than on water-treated controls. Many ofthe larvae quickly ceased feeding on treated mate-rial. Larvae took 18 min to penetrate the control discsand caused 75% damage (using a modification of

Vilardebo’s (1973) CI) compared to disks treated withNSP (55 min; 19%); NKP (162 min; 15%); NC (25 min;22%); NO (34 min, 7%).

Larval growth: After 4 and 14 days, respectively, mor-tality of second-instar larvae was statistically similar oncontrols (17%; 17%) than on treated material (25–40%;40–60%). However, larval weight was 297 mg in con-trols, 188 mg in NC, and 61–81 mg in other treatments.

Population build-up and damage: Three months afterrelease of weevils at the base of bananas planted indrums, populations were 54–270% higher in controlsthan where NC or NSP had been integrated into thesoil. In additional drum experiments, weevil damagein controls was greater than in neem-treated materi-als, even though adult populations were similar amongtreatments.

Phytotoxicity: NKP and NO appeared toxic to thebanana plants and may have interfered with nutri-ent and water uptake. NSP and NC displayed phyto-toxic effects only at application rates >100 g/plant.Phytotoxicity levels may have been affected by soiltype and azadirachtin content (found to be high in theNC used in these trials).

Application methods, frequency, rates: Direct appli-cation of NC and NSP to the soil was much morecost-effective than applications of aqueous solutions.Overall, NSP appeared to be the preferred deriva-tive as it was easier to produce and had bettereffects. From these results, Musabyimana (1999) rec-ommended application rates was 60–100 kg/ha onceevery 4 months.

Field trials: In field trials at 3 sites, applications ofNC and NSP (1) contributed to higher sucker establish-ment; (2) had little impact on adult numbers; (3) pro-vided major reductions in weevil damage; (4) reducednematode damage; and (5) contributed to yield advan-tages of up to 50% across 2 crop cycles. However, onlyhigh application rates reduced weevil damage at thesite where weevil damage was most heavy. This appli-cation rate also resulted in phytotoxicity problems andloss of the ratoon crop.

Musabyimana’s (1999) results suggest that neemderivatives can reduce weevil damage by interferingwith each stage of attack: (1) fewer adults will locateor remain at the host plant; (2) females locating the hostplant will have reduced oviposition; (3) eclosion rates


will be reduced; (4) an antifeedant effect will delayand reduce larval feeding; and (5) larval fitness willbe reduced. The influence of neem applications on lar-val success suggest that neem derivatives also have asystemic effect.

In laboratory studies in Cameroon, Messiaen(2000, 2002) had results consistent with those ofMusabyimana (1999): Neem had a repellent effecton adults and reduced oviposition levels and eclosionrates. Braimah (1997) found that potted soil whichhad previously supported neem plants were repellentto weevils. Silva & Fancelli (1998) also found neemto be a repellent to adults but provided little detail.In addition, Messiaen (2000, 2002) reported that neemaffected fertility of female weevils.

However, in field studies, Messiaen et al. (2000) andMessiaen (2002) found limited advantage in weevilcontrol from applications of neem dips (i.e. aqueoussolution of concentrated NSP) and no benefits fromgranular applications of NSP (30–100 g/plant). Neemtreatments did not have any effect on weevil adult pop-ulations in either of two trials. However, neem dipsreduced sucker mortality by 73–85% and total plantmortality by 50%. Neem dips also reduced damage inone trial by 70% but had no effect on weevil damagein a second trial. In contrast, had no effect on plantmortality or weevil damage.

From his results in Kenya, Musabyimana (1999)concluded that NSP and NC soil applications are effec-tive enough to do away with paring and hot watertreatment of suckers to be used as planting material.His data suggest that extended protection under fieldconditions is possible. However, the largely negativeresults obtained by Messiaen et al. (2000) and Messiaen(2002) in Cameroon were inconsistent with those ofMusabyimana (1999) and show that it would be usefulto conduct further studies at additional sites. In addi-tion, the availability of neem products, their economicviability and their acceptance by farmers in differentbanana production systems needs to be determined.

XVI. Pesticides

Chemical pesticides for control of banana weevil maybe applied to protect planting material (through dippingof suckers or applications in planting holes), periodi-cally applied at the base of the mat after crop establish-ment, and/or applied to pseudostem traps to increasetrap catches. Since the first recommendation in 1907for the use of chemicals to control banana weevil,

there have been numerous studies on the relativeefficacy of different insecticides under different for-mulations and application rates, persistence, and theappearance of insecticide resistance in banana weevils.Chemicals remain an important part of banana weevilcontrol although costs often make them prohibitive forsubsistence farmers.

The early use of non-synthetic pesticides againstbanana weevil has been reviewed by Viswanath (1976).Gravier (1907) recommended immersing suckers inBordeau mixture. During the next 20 years, a range ofchemicals, including sodium arsenite, mercuric chlo-ride, lead arsenate, Paris green, calcium arsenate, andborax were tested against the weevil. Of these Parisgreen and sodium arsenite were deemed the most effec-tive (Froggatt 1924, 1925; Veitch 1929; Sein 1934;Weddell 1945).

In 1951, the use of chemicals gained further impor-tance with the advent of synthetic insecticides thatlargely replaced labour-demanding cultural controlssuch as trapping or sanitation (Braithwaite 1958;Vilardebo 1984; Simon 1994). As with many otherpests, the introduction of chemicals in the 1950s wasgreeted with optimism. Braithwaite (1958) suggestedthat eradication of the banana weevil might be achievedwith aldrin and dieldrin.

Since the introduction of synthetic insecticides, awide range of chemicals, encompassing all of themajor classes, have been tested and recommendedas effective for the control of banana weevil (reviewed,in part, by Sponagel et al. (1995) and Seshu Reddyet al. (1998)). These have include aldicarb, aldrin, ben-diocarb, cadusafos, carbaryl, carbofuran, carbosulfan,chlordane, chlorfenvinphos, chlorpyrifos, chlotecore,cyfluthrin, DDT, diclorvos, dieldrin, dimethoate, disul-foton, ekadrin, endosulfan, endrin, EPN, ethoprop,fensulfothion, fenthion, fipronil, HCH, heptachlor,isazofos, isofenphos, kepone, lindane, mephos-folan, monocrotophos, omethoate, oxamyl, parathion,phenamiphos, phorate, pirimiphos-ethyl, profenofos,propoxur, prothiophos, tebupirimphos, triazophos, andtrichlorphon. Some of these chemicals (e.g. isazophos,oxamyl, phenamiphos) served primarily as nematicidesbut also provided good control against banana weevil(Robalino et al. 1983; Bujulu et al. 1986). Recommen-dations include sucker drenches, and applications toplanting holes, the base of the mat and to pseudostemtraps. Many previously recommended chemicals havesince been banned or otherwise fallen out of favourfor high levels of mammalian toxicity, environmentalconcerns, and/or the development of resistance.


Chemical pesticides tend to be more regularlyused in commercial plantations, while insecticide useis much less for low-resource, subsistence growers.During rapid rural appraisals at 25 sites in Uganda in1991, for example, few farmers reported use of chem-ical insecticides in banana fields (Gold et al. 1993).Most farmers claimed that they could not afford, hadno access to or no information on how to use insecti-cides. At seven sites, a majority of farmers expressed adesire to apply chemicals against banana weevil, if theywere subsidised or made more affordable. Many farm-ers found the use of insecticides against banana weevilappealing because chemicals require little labour, arefast-acting and appear to be a reliable means of con-trol. In contrast, farmers had more limited confidencein cultural control methods which require labour inputsand for which results might not be apparent for manymonths.

In a commercial growing region of Masaka dis-trict, Uganda an estimated 30–40% of the farmers inMasaka regularly used pesticides (70% carbofuran) tocontrol banana weevil (Gold et al. 1999a). Elsewherein the district, however, less than 30% of farmersapplied chemicals (mostly carbofuran) against theweevil (Ssennyonga et al. 1999). An equal propor-tion of farmers had abandoned the use of insec-ticides, primarily because of cost. Those farmerswho continued to use insecticides tended to becommercial farmers and in the upper economic strataof the community. Most of those who used carbo-furan reported it to be very effective at controllingweevils.

Insecticide resistance in banana weevil has been doc-umented in Australia (Kelly 1966; Vilardebo 1967;Swaine & Corcoran 1973; Edge 1974; Shanahan &Goodyer 1974; Edge et al. 1975; Wright 1977; Swaineet al. 1980; Collins et al. 1991), Latin America(Sotomayor 1972; Foreman 1976; Mitchell 1978;Mello et al. 1979; Sampaio et al. 1982) and Africa(Bujulu et al. 1983; Gold et al. 1999a) for a rangeof chemicals including cyclodienes (aldrin, BHC,heptachlor, dieldrin), organophosphates (chlorpyrifos,ethoprophos, pirimiphos-ethyl, and prothiophos) andcarbamates (carbofuran). Cross-resistance has alsobeen demonstrated (Edge 1974; Collins et al. 1991).Castrillon (2000) suggests that in concert with thedevelopment of resistance, pesticides upset naturalcontrol by endemic natural enemies, leading to greaterweevil pressure. Sengooba (1986) and Sebasigari &Stover (1988) attributed weevil outbreaks in Ugandato the development of resistance to dieldrin.

Roberts (1958) attributed outbreaks of anotherbanana weevil, M. hemipterus, to applications of dield-rin which he believed eliminated natural enemies,including ants. In Uganda and Tanzania, outbreaksof banana weevil in the mid-1980s were attributed topest resurgence following development of resistance todieldrin (Sengooba 1986; Sebasigari & Stover 1988;Taylor 1991; Gold et al. 1999a) leading to loss ofconfidence in chemical control by some farmers.

XVII. Summary and Conclusions

Bananas and plantains are important cash and sub-sistence food crops throughout the tropics and sub-tropics. Banana is a genetically diverse crop withnumerous clones (including diploids, triploids, andtetraploids) that may be grown under highly dis-parate cropping and management systems. Banana isgrown from sea level to >2000 masl. Production sys-tems range from low-input kitchen gardens and smallstands to intensive, large-scale commercial bananaplantations serving export markets in Europe and NorthAmerica. Small-scale production is often extremelyimportant in the livelihoods of many third-world farm-ers. In Africa, Latin America, and Asia, a wide vari-ety of clones (including dessert, cooking, roasting, andbrewing types) serve local markets and contribute tothe food security of the rural poor. The highland cook-ing banana is the primary staple in the East AfricanGreat Lakes region, while plantains are important foodsthroughout West Africa and Latin America.

The banana weevil is the most important insect peston bananas and plantains. Studies of banana weevilbegan in the early 1900s, although most research hasbeen conducted since 1980. Much of the informationon the weevil has been published in theses, proceedingsand reports. In some cases, it is hard to separate con-clusions and recommendations based on the author’sdirect observations and experiences, as opposed to reit-eration of what has already been written elsewhere.Recommendations to farmers have often been based oncasual field observations, suppositions and hypothesesthat have not been supported by scientific evidence.

Research findings are often hard to interpret. InUganda, for example, damage levels showed only aweak relationship with adult densities, while popu-lation shifts did not relate well with the number ofweevils removed through systematic trapping. Surveysin Uganda have demonstrated wide variability inweevil populations and damage on adjacent farms


in similar environments, suggesting that managementhas a strong influence on weevil pest status. Analysisof data, however, did not provide clear relationshipsbetween most management parameters and damagealthough the most important factor appeared to be cropsanitation (Gold et al. 1997, unpubl. data).

Data collected from different sites are often contra-dictory. The weevil has been variously reported to bemost active in the dry season, the wet season or dis-play activity patterns independent of weather factors.Variability in larval developmental rates has also beenreported by different researchers working at proximalsites. Studies on cultivar susceptibility have also pro-duced variable results; reports of weevil pest statuson Cavendish, for example, range from unimportantto very serious. The inconsistency in research find-ings across studies may reflect differences in bananaclones, management and production systems, agro-ecological conditions, weevil biotypes, and researchmethodologies.

Nevertheless, certain aspects of the banana weevil’sbiology appear clear. The banana weevil is charac-terised by nocturnal activity, long life span, limitedmobility, low fecundity, and slow population growth.The sex ratio is 1 : 1. The adults are free living and mostoften associated with banana mats and cut residues.Flight is rare or uncommon and movement by crawl-ing is limited. Dissemination is primarily through themovement of infested planting material. The weevilis attracted to their hosts over short distances byvolatiles. Cut corms, including recently detached suck-ers used as planting material, are especially attrac-tive. Males produce an aggregation pheromone that isresponded to by both sexes.

The adults often live more than 1 year, but pro-duce only a few eggs per week. Oviposition is inthe corm or lower pseudostem. The immature stagesare passed within the host plant, mostly in the corm.Developmental periods in degree-days have been deter-mined for the different immature stages. Under ambientconditions, the egg stage lasts 1 week, the larval stageabout 1 month and the pupal stage 1 week. Populationbuild-up is slow. The weevil displays only weak den-sity dependence effects, suggesting that high mortal-ity to the immatures acts as a brake to populationgrowth.

Highland banana and plantains are particularly sus-ceptible to banana weevil damage. In East Africa,weevil attack has led to accelerated yield declines inmuch of the region and the replacement of highlandbananas with exotic brewing bananas that are resistant

to this pest. Severe weevil problems may also appearin certain regions on clones commonly perceived asresistant (e.g. Cavendish in South Africa), suggest-ing that environment may also play a role in deter-mining weevil. Data on yield loss to banana weevilare limited. It is clear, however, that weevil problemsbecome increasingly severe in ratoon crops, althoughthe weevil can sometimes be a serious problem in newlyplanted stands. The weevil may extend the length ofthe crop cycle, cause reductions in bunch weight andcontribute to plant loss through toppling and snapping.Mat die-out and shortened plantation life have also beenattributed to weevil attack.

The weevil’s biology creates sampling problems andmakes its control difficult. Most commonly, weevilsare monitored by trapping adults, mark and recapturemethods, and damage assessment to harvested or deadplants. A range of sampling methods have been pro-posed to assess damage, of which the most impor-tant have been the CI (Vilardebo 1973), PCI (Mitchell1980), and cross section estimates of damage to thecentral cylinder and cortex (Gold et al. 1994b). Allof these require destructive sampling and are oftensubjective, making comparisons between studies dif-ficult. Estimates of damage to the central cylinder and,possibly cortex, may best reflect the impact of theweevil on plant growth and yield (Rukazambuga 1996).Establishing agreed upon sampling protocols should bea high priority among banana weevil researchers.

Research results suggest that no single method willbring about complete control of the banana weevil andthat there is no ‘silver bullet’ waiting to be found.Therefore, a broad IPM strategy appears to be the bestapproach to addressing banana weevil problems. Thisincludes cultural control, biological and microbial con-trol, host plant resistance, the use of botanicals, andchemical control. Adoption of different components islikely to be affected by the farmer’s perception of theimportance of the weevil and his level of resources.

Cultural controls of banana weevil have been widelypromoted and are available to most farmers. The mostimportant of these methods are the use of clean plant-ing material, crop sanitation and agronomic meth-ods to improve plant vigour and tolerance to weevilattack, neem, and trapping. A combination of thesemethods is likely to provide at least partial controlof banana weevil. However, all of these methods havecosts and adoption by resource-poor subsistence farm-ers is often limited. Moreover, few controlled studieshave been undertaken to demonstrate the benefits ofthese methods.


The use of clean planting material is important inexcluding banana weevils and other pests from newlyplanted banana stands. Tissue culture plantlets are rou-tinely used in commercial banana production through-out the world and are being promoted for subsistencefarmers in some countries. Access to tissue cultureand associated costs are limiting factors for dissemi-nation of this technology. Other methods (e.g. paring,hot water treatment) of cleaning planting propagulesare also available where access to tissue culture mate-rial is not possible. Paring requires little labour onthe part of the farmer. In contrast, hot water treat-ment is often good in theory but very difficult in prac-tice because of material requirements. Moreover, underconditions of land pressure, many banana stands areplanted in or proximal to previously infested fields. Insuch cases, re-infestation is an important concern andthe use of clean planting material is not a long-termsolution to the banana weevil problem.

Crop sanitation and agronomic practices to pro-mote plant vigour and tolerance to weevil attackappear to be common-sense approaches to the weevilproblem. These methods are widely recommendedalthough few data are available to show that theyreduce weevil pressure. Although the employment ofhigh standards of agronomic practices in maintaningstand productivity can not be disputed, their role inbanana weevil control is not clear. In an on-stationtrial, Rukazambuga et al. (2002) demonstrated greateryield losses (tonnes/ha) in vigorously-growing bananathan in stressed systems, while McIntyre et al. (2002)concluded that weevil and nematode-infestation inestablished banana fields impeded uptake of nutri-ent amendments. Musabyiamana (1999) successfullyreduced banana weevils through applications of neem.Further testing on the use of neem against bananaweevil should be undertaken.

Trapping of banana weevils with pseudostem trapsand disk-on-stump traps has been widely promoted,although the overall benefit of trapping has been con-troversial. Gold et al. (2002b) demonstrated throughfarmer-participatory research trials that although sys-tematic pseudostem trapping reduced banana weevilpopulations on most farms, subsequent farmer adop-tion was low due to labour and material requirements.Enhanced trapping with synthetic pheromone luresand kairomones is rightfully a priority for currentstudy.

Between 1912 and 1938, researchers exploredthe prospects for classical biological control ofbanana weevil. Generalist, opportunistic predators

were collected in Indonesia and released in the Pacific,Africa, and Latin America. These predators eitherdid not establish or failed to bring about control(Waterhouse & Norris 1987). Recent searches forbanana weevil parasitoids in Indonesia had nega-tive results. Additional searches in India, the centreof origin of weevil-susceptible plantains, are prob-ably warranted. Biological control using ants maybe possible (Roche & Abreu 1982, 1983), but theefficacy of other endemic predators seems limited(Koppenhofer & Schmutterer 1993).

Microbial control may offer promise for the con-trol of banana weevil. Numerous strains of B. bassianaand M. anisopliae have been demonstrated to killhigh percentages (i.e. >95%) of banana weevils whenapplied topically to adults in the laboratory. To date,research has focused on (1) surveys of naturally occur-ring infections; (2) pathogenicity studies comparingstrains, spore concentrations, formulations and modesof application; (3) fungal production on different sub-strates; and (4) a limited amount of field testing on fun-gal persistence and population suppression (Godonou1999; Nankinga 1999). Entomopathogenic nematodeshave also been shown to be cause high levels of mor-tality banana weevils in both the laboratory and field(Treverrow et al. 1991; Schmitt et al. 1992). Currentresearch priorities should include the development ofeconomic mass production and delivery systems andevaluation of fungal performance and efficacy underdifferent agro-ecological conditions. Unless these aredeveloped, the use of entomopathogenic fungi andnematodes as biopesticides will either be beyond themeans of most farmers or not competitive with the costsof chemical insecticides.

The use of endophytes for control of banana weevilmay also be possible (Griesbach 1999), althoughresearch in this area is still in its relative infancy.

Breeding efforts in banana have focused on develop-ing resistance to nematodes and diseases. To date, therehave been no attempts to breed for resistance to bananaweevil. More recently, however, there has been increas-ing recognition of the importance of banana weevilas a worldwide problem (Anonymous 2000). At thesame time, there have been advances in both conven-tional and non-conventional breeding of banana thatmay offer exciting opportunities for developing weevil-resistant hybrids. Screening trials have demonstratedthe availability of many clones, including Calcutta-4,Yangambi-km5, and FHIA-03, that are resistant tothe banana weevil and might be utilised in breedingprogrammes. Antibiosis appears to be predominant


mechanism conferring resistance to weevils withinMusa germplasm.

Biotechnological approaches, including genetictransformation, might facilitate the development ofweevil-resistant clones that retain many of the locallydesirable fruit characteristics. For example, it may bepossible to identify candidate genes from within andwithout the Musa genome and through currently avail-able transformation systems, study the expression ofsuch genes in banana. This could include both damage-induced gene expression (i.e. in known resistant Musacultivars) and enzyme (amylase and protease) inhibitorgene expression, using foreign genes of plant origin.

In summary, available cultural controls may con-tribute to suppressing populations of banana weevil,but are unlikely to offer complete control in stands ofhighly susceptible germplasm or regions where pestpressure is high. Chemicals often offer complete con-trol, but their costs, the development of weevil resis-tance against all classes of insecticides and ecologicalside effects mitigate against the use of chemical controlas a long-term strategy.

The way forward appears to be through the improvedmanagement in small farmer systems, the refinementof microbial control mass production and delivery sys-tems and the development of both conventional andnon-conventional breeding for host-plant resistance.Further studies on the use of some endemic natural ene-mies (e.g. myrmicine ants), the use of semiochemical-based trapping systems and botanicals, such as neem,also appear to be warranted.

Acknowledgements

We thank Caroline Nankinga (NARO, Uganda),Andrew Kiggundu (NARO, Uganda) and two anony-mous reviewers for their critical comments on earlierdrafts of this paper. We are grateful to the follow-ing people for their personal communications and theuse of unpublished data: Agnes Abera (IITA-ESARC,Kampala, Uganda), Ignace Godonou (CABI, Nairobi),Ahsol Hasyim (Research Institute for Fruits, Solok,Indonesia), Godfrey Kagezi (IITA-ESARC), SlawomirLux (ICIPE, Nairobi, Kenya), Michael Masanza(IITA-ESARC), Beverly McIntyre (formerly NARO,Uganda), Gertrude Night (IITA-ESARC), VincentOchieng (ICIPE), Cam Oehlschlager (Chemtica,International, San Jose, Costa Rica), Suleman Okech(IITA-ESARC), Eugene Ostmark (formerly FHIA,La Lima, Honduras), S. Rodriguez (INIVIT, Santa

Clara, Cuba), Daniel Rukazambuga, (formerly Ministryof Agriculture, Tanzania), K.V. Seshu Reddy (ICIPE),Henry Ssali (NARO, Uganda), William Tinzaara(NARO, Uganda), Lancine Traore (formerly IITAand Katholiecke University, Leuven), Altus Viljoen(University of Pretoria, South Africa). We alsowish to acknowledge the support of WilberforceTushemereirwe (NARO, Uganda), the late PaulSpeijer (IITA-ESARC), Peter Neuenschwander (IITA,Cotonou), John Lynam (Rockefeller Foundation,Nairobi), Andrew Kerr and Luis Navarro (IDRC,Nairobi) and the University of Florida TropicalResearch and Education Center for their support andencouragement in carrying out our own research onbanana weevils. We thank Claudine Picq (INIBAP,Montpellier) for her assistance in obtaining hard toretrieve literature.

References cited

Abera, A.M.K. (1997) Oviposition Preferences and Timing ofAttack by the Banana Weevil (Cosmopolites sordidus Germar)in East African Highland Banana (Musa spp), 120 pp. Mastersthesis, Makerere University, Kampala, Uganda.

Abera, A.M.K., Gold, C.S. and Kyamanywa, S. (1999) Timingand distribution of attack by the banana weevil (Coleoptera:Curculionidae) in East African highland banana (Musa spp.)Fla. Entomol. 82, 61–641.

Afreh-Nuamah, K. (1993) Population dynamics of Cosmopolitessordidus in relation to sources of planting material and crop-ping history at Kade, Ghana. In C.S. Gold and B. Gemmill(eds) Biological and Integrated Control of Highland Bananaand Plantain Pests and Diseases. Proceedings Research Coor-dination Meeting, pp. 68–74. Cotonou, Benin: IITA.

Aguero, J.V. (1976) Siembra de platanos y cambures libresde Cosmopolites sordidus y de nematodos (Radopholussimilis, Pratylenchus, Helicotylenchus y Meloidogyne). BoletinInformativo Ministerio de Agricultura y Cria 5, 1–3.

Ahiekpor, E.K.S. (1996) Plantains in Ghana: A brief synopsis.In R. Ortiz and M.O. Akoroda (eds) Plantain and BananaProduction and Research in West Africa: Proceedings of aRegional Workshop. 23–27 September 1995, pp. 43–4. Ibadan,Nigeria: IITA.

Allen, R.N. (1989) Control of Major Pests and Diseases ofBananas, Information from the Department of Agriculture,New South Wales, 10 pp.

Alpizar, D., Fallas, M., Oehlschlager, A.C., Gonzalez, L. andJayaraman, S. (1999) Pheromone-based mass trapping ofthe banana weevil, Cosmopolites sordidus (German) and theWest Indian sugarcane weevil Metamasius hemipterus L.(Coleoptera:Curculionidae) in plantain and banana. InMemorias XIII Reunion ACORBAT, 23–27 November 1998,pp. 515–38. Guayaquil, Ecuador.

Altieri, M.A. and Letourneau, D.K. (1982) Vegetation manage-ment and biological control in agroecosystems. Crop Prot. 1,405–30.


Altre, J.A. and Vandenberg, J.D. (2001) Factors influencing theinfectivity of isolates of Paecilomyces fumosoroseus againstdiamondback moth, Plutella xylostella. J. Invertebr. Pathol. 78,31–6.

Ambrose, E. (1984) Research and development in bananacrop protection (excluding Sigatoka) in the English speakingCaribbean. Fruits 39, 234–47.

Anitha, N., Rajamony, L. and Radhakrishnan, T.C. (1996)Reaction of banana clones against major biotic stresses. Planter72, 315–21.

Anonymous (1989) Guia Educativa No. 1: Trampeo para elpicudo negro en platano, FHIA, La Lima, Honduras, 14 pp.

Anonymous (1992) Mejoramiento del cultivo del platano enla zona cafeteria de Colombia. FNCC-Cenicafe, ICA andIRFA-CIRAD, Colombia, 54 pp + annexes.

Anonymous (2000) Banana weevils as new PROMUSA priority.PROMUSA Supplement 6, iii–iv in Infomusa 9(2).

Aranda, O.A. (1976) Evaluacion del dano causado por elpicudo negro del platano Cosmopolites sordidus Germ.(Coleoptera:Curculionidae) en la Chontalpa, Tab. In Memorias:Simposio Nacional de Parasitologia Agricola, Mexico, MexicoD.F. Ing. Agron. Parasitol. 4, 165–78.

Aranzazu, L.F., Arcila, M.I., Bolanos, M.M., Castellanos, P.A.,Castrillon, C., Perez, J.C., Rodriguez, J.L. and Balencia. J.A.(2000) Manejo Integrado del Cultivo de Platano. ManualTecnico. CORPOICA, Manizales, Colombia, 80 pp.

Aranzazu, L.F., Munoz, C.I., Castellanos, P.A., Castrillon, C.,Bolanos, M.M., Arcila, M.I., Valencia, J.A., Perez, J.C.,Rodriguez, J.L., Lucas, J.C. and Diaz, L.B. (2001)Capacitacion y transferencia de tecnologia para con-tribuir al mejoramiento del agronegocio del platano en losDepartamentos del Quindio y Valle del Cauca. CORPOICA,Manizales, Colombia, 130 pp.

Arleu, R.J. (1982) Dinamica populacional e controle doCosmopolites sordidus (Germ., 1824) e Metamasius hemipterusL., 1764 (Col.: Curculionidae), em bananais da cv. Prata, noEspirito Santo, Brasil. Piracicaba. ESALQ. 66 pp.

Arleu, R.J. (1983) Broca da bananeira Cosmopolites sordidus(Germ., 1824) Coleoptera-Curculionidae na cultivar Prata. InSimposio sobre Bananeira Prata, 1, Cariacica, Espirto Santo,7–11 November 1983, pp. 36–45.

Arleu, R.J. and Neto, S.S. (1984) Broca da bananeiraCosmopolites sordidus (Germ., 1824) (Coleoptera:Curculionidae). Turrialba 34, 359–67.

Arleu, R.J., Neto, S.S., Gomes, J.A., Nobrega, A.C. andSardini, D.M. (1984) Dinamica populacional do Cosmopolitessordidus (Germ., 1824) (Col.:Curculionidae) em bananais dacv. Prata (Grupo AAB), em Alfredo Chaves, Espirito Santo.Turrialba 34, 473–80.

Arroyave, F.P. (1985) Control del picudo negro Cosmopolitessordidus Germar en semilla vegatativa de platano (Musa AABSimmonds), 112 pp. Tesis, Ing. Agr. Universidad de Caldas,Colombia.

Arthurs, S. and Thomas, M.B. (2001) Effects of temperatureand relative humidity on sporulation of Metarhizium ansiopliavar. acridum in mycosed cadavers of Schistocerca gregaria.J. Invertebr. Pathol. 78, 59–65.

Ayala, J.L. and Monzon, S. (1977) Ensayo sobre diferentes dosisde Beauveria bassiana para el control del picudo negro delplatano (Cosmopolites sordidus) (Germar). Centro Agric., Rev.Cien. de Fac. Cienc. Agric. 4, 19–24.

Bakyalire, R. (1992) A study of the life cycle and behaviour ofthe banana weevil Cosmopolites sordidus Germar, in Uganda,118 pp. Masters thesis Makerere University, Uganda.

Bakyalire, R. and Ogenga-Latigo, M.W. (1992) Aspectsof the life cycle and behavior of the banana weevil,Cosmopolites sordidus Germar (Coleoptera: Curculionidae).In M.W. Ogenga-Latigo (ed) Recent Contribution to BananaEntomology in Uganda (1990–1992), pp. 6–19. Kampala,Uganda: Department of Crop Sciences. Makerere University.

Barrera, J.F. and Jimenez, E. (1994) Establecimiento de Plaesiusjavanus (Coleoptera:Histeridae) en Chiapas, Mexico para elcontrol de Cosmopolites sordidus (Coleoptera:Curculionidae).Vedalia 1, 23–4.

Barriga, R. and Montoya, R. (1972) Seleccion de semillas debanano y platano. Bol. Agric. (Medellin) 623, 12874–9 (citedin Arroyave 1985).

Batista Filho, A., Paiva Castro, L.M., Myazaki, I., BastosCruz, B.P. and Oliveira, D.A. (1987) Controle biolgico do‘moleque’ da bananeira Cosmopolites sordidus, Germar, 1824)pelo uso de fungo entomogenos, no laboratorio. Biologico (SaoPaulo) 53, 1–6

Batista Filho, A., Leite, L.G., Raga, A. and Sato, M.E. (1990).Atracao de Cosmopolites sordidus Germar (Coleoptera:Curculionidae) por iscas do tipo ‘sanduiche’ e ‘telha’. Arq.Inst. Biolog. (Sao Paulo) 57, 9–13.

Batista Filho, A., Sata, M.E., Leite, L.G., Raga, A. andPrada, W.A. (1991) Utilizacao de Beauveria bassiana (Bals.)Vuill., no controle do moleque da bananeira Cosmopolitessordidus Germar, 1824 (Coleoptera: Curculionidae). Rev. Bras.Frutic. (Cruz das Almas) 13, 35–40.

Batista Filho, A., Leite, L.G., Sato, M.E. and Raga, A. (1992)Cosmopolites sordidus (Germar, 1824) em dois cultivaresde banana: Nivel de infestacao e incidencia natural doentomopatogeno Beauveria amorpha (Hohn). Rev. Agric.(Piracicaba) 67, 183–90.

Batista Filho, A., Leitao, A.E.F., Sato, M.E., Leite, L.G.and Raga, A. (1994) Efeito da associacao Beauveriabassiana (Bals.) Vuill. com oelo mineral, na mortalidade deCosmopolites sordidus Germar (Coleoptera: Curculionidae).An. Soc. Entomol. Brasil 23, 379–83.

Batista Filho, A., Leite, L.G., Raga, A. and Sato, M. (1995a)Enhanced activity of Beauveria bassiana associated with min-eral oil against Cosmopolites sordidus (Germar) adults. An.Soc. Entomol. Brasil. 24, 405–8.

Batista Filho, A., Leite, L.G., Raga, A., Sato, M.E. andOliveira, J.A. (1995b) Utilizacao de Beauveria bassiana (Bals.)Vuill. no manejo de Cosmopolites sordidus Germar, 1824, emMiracatu, SP. Biologico 57, 17–9.

Batista Filho, A., Leite, L.G., Alves, E.B. and Aguiar, J.C.(1996) Controle de Cosmopolites sordidus (Coleoptera:Curculionidae) por fipronile e seu efeito sobre Beauveriabassiana. Arq. Inst. Biol. 63, 47–51.

Beauhaire, J., Ducrot, P.H., Malosse, C., Rochat, D., Ndiege, I.O.and Otieno, D.O. (1995) Identification and synthesis of sor-didin, a male pheromone emitted by Cosmopolites sordidus.Tetrahedron Lett. 36, 1043–6.

Beccari, F. (1967) Contributo alla conoseenza del Cosmopolitessordidus Ger. (Coleoptera, Curculionidae), Parte I–II. Riv.Agric. Subtrop. Trop. 61, 51–93; 131–50.

Bendicho, A. (1987) Poder de percepcion de la hormigaTetramorium guineense para el control biologico del picudonegro del platano. Cienc. Agric. 30, 13–5.


Bendicho, A. and Gonzales, N. (1986) Comportamiento de pobla-ciones de Cosmopolites sordidus y Tetramorium guineense encondiciones naturales. Cienc. Agric. 17, 9–12.

Boivin, G. (1993) Les parasitoides des oeufs de Curculionidae.In C.S. Gold and B. Gemmill (eds) Biological and IntegratedControl of Highland Banana and Plantain Pests and Diseases.Proceedings of a Research Coordination Meeting, pp. 97–106.Cotonou, Benin: IITA.

Boscan de Martinez, N. and Godoy, F. (1989) Epocas de inciden-cia de Cosmopolites sordidus G. y de Metamasius hemipterusL. en dos huertos de musaceas en el estado de Aragua. Agron.Trop. 38, 107–19.

Bosch, C., Lorkeers, A., Ndile, M.R. and Sentozi, E. (1996)Diagnostic Survey: Constraints to Banana Productivity inBukoba and Muleba Districts, Kagera region, Tanzania.Tanzania/Netherlands Farming Systems Research Project/LakeZone. Working Paper No. 8. Agricultural Research Institute,Maruka, Tanzania, 10 chapters + appendices.

Braimah, H. (1997) Laboratory Studies on the Host PlantSearching Behaviour and Chemical Ecology of the BananaWeevil, Cosmopolites sordidus (Germar 1824), (Coleoptera:Curculionidae), 311 pp. Ph.D. thesis, University of Reading,UK.

Braimah, H. and van Emden, H.F. (1999) Evidence for the pres-ence of chemicals attractive to the banana weevil, Cosmopolitessordidus (Coleoptera: Curculionidae) in dead leaves. Bull.Entomol. Res. 89, 485–91.

Braithwaite, B.M. (1958) Ground spray treatments for controlof the banana beetle borer (Cosmopolites sordidus) (Germar).J. Aust. Inst. Agric. Sci. 24, 27–34.

Braithwaite, B.M. (1967) Banana beetle borer control investiga-tions on the north coast of New South Wales. Agric. Gaz. NSW78, 359–65.

Breen, J.P. (1994) Acremonium endophyte interactions with plantresistance to insects. Ann. Rev. Entomol. 39, 401–23.

Brenes, S. and Carballo, M. (1994) Evaluacion de Beauveriabassiana (Bals.) para el control biologico del picudo del pla-tano Cosmopolites sordidus Germar. Manej. Integr. Plagas 31,17–21.

Bridge, J. and Gowen, S.R. (1993) Visual assessment of plant par-asitic nematodes and weevil damage on bananas and plantain.In C.S. Gold and B. Gemmill (eds) Biological and IntegratedControl of Highland Banana and Plantain Pests and Diseases.Proceedings of a Research Coordination Meeting, pp. 147–54.Cotonou, Benin: IITA.

Budenberg, W.J. and Ndiege, I.O. (1993) Volatile semiochem-icals of the banana weevil. In C.S. Gold and B. Gemmill(eds) Biological and Integrated Control of Highland Bananaand Plantain Pests and Diseases. Proceedings of a ResearchCoordination Meeting, pp. 75–86. Cotonou, Benin: IITA.

Budenberg, W.J., Ndiege, I.O. and Karago, F.W. (1993a) Evidencefor volatile male-produced pheromone in banana weevilCosmopolites sordidus. J. Chem. Ecol. 19, 1905–15.

Budenberg, W.J., Ndiege, I.O., Karago, F.W. and Hansson B.S.(1993b) Behavioral and electro-physiological responses on thebanana weevil Cosmopolites sordidus to host plant volatiles.J. Chem. Ecol. 19, 267–77.

Bujulu, J., Uronu, B. and Cumming, C.N.C. (1983) The controlof banana weevils and parasitic nematodes in Tanzania. EastAfr. Agric. For. J. 49, 1–13.

Bullock, R. and Evers, C. (1962) Control of the banana rootborer (Cosmopolites sordidus Germar) with granular insecti-cides. Trop. Agric. 39, 109–13.

Busoli, A.C., Fernandes, O.A. and Tayra, O. (1989) Controleda broca da bananeira Cosmopolites sordidus Germar 1824(Coleoptera, Curculionidae) atraves dos fungos entomopato-genicos Beauveria bassiana (Bals) Vuill. e Metarhiziumanisoplae (Metschn.) Sorok. (Hyphomycetes). An. Soc. Ent.Brasil 18(Suppl.), 33–41.

Calderon, A., Castineiras. A. and Lopez, M. (1991) Efecto de losbiocidas y fertilizantes empleados en el cultivo del platano enCuba sobre los hongos entomopatogenos. Prot. Plant. 1, 21–31.

Carballo, M. (1998) Mortalidad de Cosmopolites sordidus condiferentes formulaciones de Beauveria bassiana. Manej. Integr.Plagas 48, 45–8.

Carballo, M. and de Lopez, M.A. (1994) Evaluacion de Beauveriabassiana (Bals.) para el control biologico del Cosmopolitessordidus Germar y Metamasius hemipterus en condiciones decampo. Manej. Integr. Plagas 31, 22–4.

Cardenas, R. (1983) El picudo negro del platano: Cosmopolitessordidus (Germar). In 1er Seminario Internacional sobrePlatano, 1. Manizales, Colombia, Memorias, 6–10 June 1983,pp. 128–134. Universidad de Caldas, Manizales, Colombia.

Cardenas, R. and Arango, L.G. (1986) Fluctuacion poblacional ydispersion del picudo negro del platano Cosmopolites sordidus(Germar 1824). Rev. Colomb. Entomol. 12, 37–45.

Cardenas, R. and Arango, L.G. (1987) Control del picudo negroCosmopolites sordidus (Germar 1824) del platano Musa AAB(Simmonds) mediante practicas culturales. Cenicafe 38, 50–61.

Carnero, A., Padilla, A. and Montesdeoca, M. (2002)Metodos alternativos para el control del picudo de la pla-tanera Cosmopolites sordidus Germar, 1824 (Coleoptera:Curculionidae). In D. Fernandez and P.M. Hernandez (eds)Actividades del ICIA en Platanera. Instituto Canario de Ivesti-gaciones Agrarias, pp. 75–81. Tenerife, Canary Islands, Spain.

Carroll, G.C. (1991) Fungal associates of woody plants as insectantagonists in leaves and stems. In P. Barbosa, V.A. Krischikand C.G. Jones (eds) Microbial Mediation of Plant–HerbivoreInteractions, pp. 253–71. New York: John Wiley and Son.

Castano, P.O. (1983) Manejo de problemas entomologicos en loscultivos de platano y banano. In Primer Seminario Interna-cional sobre el Platano. Manizales. 6–10 June 1983, pp. 8–11(cited in Arroyave 1985).

Castineiras, A. (1982) Actividad forrajera de Pheidolemegacephala (Hymenoptera:Formicidae:Myrmicinae). Cienc.Tecn. Agric. 5, 55–64.

Castineiras, A. and Ponce, E. (1991) Efectividad de la utiliza-cion de Pheidole megacephala (Hymenoptera:Formicidae) enla lucha biologica contra Cosmopolites sordidus (Coleoptera:Curculionidae). Prot. Plant. 1(2), 15–21.

Castineiras, A., Lopez, M., Calderon, A., Cabrera, T. andLujan, M. (1990) Virulencia de 17 aislamientos de Beauveriabassiana y 11 de Metarhizium anisopliae sobre adultos deCosmopolites sordidus. Cienc. Tecn. Agric. 13, 45–51.

Castrillon, C. (1987) Reconocimiento del picudo negro(Cosmopolites sordidus Germar) del platano en el Departa-mento del Quindio. ICA (Manizales) Informa 21(2), 16–21.

Castrillon, C. (1989) Plagas del cultivo del platano. In Cursode Actualizacion sobre Problemas Sanitarios en Platano.La Dorada, Colombia: ICA, 54 pp.


Castrillon, C. (1991) Manejo del picudo negro (Cosmopolitessordidus Germar) en platano y banano de la zona cafetera deColombia. ACORBAT: Mem. IX, 349–62.

Castrillon, C. (2000) Distribucion de las especies de picudo delplatano evaluacion de sus entomopatogenos nativos en el depar-tamento de Risaralda. CORPOICA, Manizales, Colombia,72 pp.

Cendana, S.M. (1922) The banana weevil. Philip. Agric. 10,367–76.

Cerda, H., Lopez, A., Fernandez, G., Sanchez, P. and Jaffe, K.(1994) Etologia y control del gorgojo negro del pla-tano Cosmopolites sordidus Germar (1824) (Coleoptera:Curculionidae) I. conducta olfactiva frente a semioquimicosde la planta huesped. ACORBAT: Mem. XI, 359–75.

Cerda, H., Lopez, A., Sanoja, O., Sancez, P. and Jaffe, K. (1995)Attraccion olfativa de Cosmopolites sordidus Germar (1824)(Coleoptera: Curculionidae) estimulado por volatiles origina-dos en Musaceas de distintas edades y variedades genomicas.Agron. Trop. 46, 413–29.

Champion, J. (1975) Productions bananieres et recherchescientifique. Fruits 30, 11–7.

Chavarria-Carvajal, J.A. (1998) Response of eight plantain clonesto nematodes and the corm-weevil (Cosmopolites sordidusGermar) in Puerto Rico. In Reunion, 23–27 November 1998,Guayaquil ACORBAT Mem. XIII, 539–46.

Coates, P.L. (1971) Effects of treatment of banana corms with asystematic nematicide. PANS 17, 448–52.

Collins, P.J., Treverrow, N.L. and Lambkin, T.M. (1991)Organophosphorous insecticide resistance and its manage-ment in the banana weevil borer, Cosmopolites sordidus(Germar) (Coleoptera:Curculionidae), in Australia. Crop Prot.10, 215–21.

Contreras, T. (1996) Evaluacion de trampas de pseudotallo y for-mulaciones de Beauveria bassiana (Bals) en el combate delpicudo negro del platano Cosmopolites sordidus (Germar) enCosta Rica. Tesis Mag. Sci. CATIE, Turrialba, Costa Rica,68 pp.

Crooker, P.S. (1979) Final Report of the ResearchOfficer/Entomology Submitted to the Director of Agriculture,Ministry of Agriculture, Fisheries and Forestry, Tonga. 32 pp.

Cuille, J. (1950) Recherches sur le charancon du bananier. Institutde Fruits et Agrumes Coloniaux. Serie Technique No. 4, Paris,225 pp.

Cuille, J. and Vilardebo, A. (1963) Les calandrini nuisibles aubananier. In A.S. Balachowsky (ed) Entomologie appliquee al’agriculture, pp. 1099–114. Masson et Cie Ed., Paris.

Davide, R.G. (1994) Status of nematode and weevil borerproblems in Philippines. In R.V. Valmayor, R.G Davide,J.M. Stanton, N.L. Treverrow and V.N. Roa (eds) Proceedingsof Banana Nematode/Borer Weevil Conf., Kuala Lumpur,18–22 April 1994, pp. 79–89. Los Banos, Philippines: INIBAP.

Dawl, N.M. (1985) Insect pest management in banana. InB. Umali and C. Lantican (eds) Proc. Inter. Seminar-WorkshopBanana Plantain Res. Dev., pp. 100–5. Los Banos, Philippines:ACIAR and PCARRD.

Deang, R., Caburubias, R. and Quero, E. (1969) Insecticide testfor the control of the abaca corm weevil. Philip. J. Plant Ind.34, 79–87.

Debach, P. (1964) Biological Control of Insect Pests and Weeds.London: Chapman and Hall, 844 pp.

Delattre, P. (1980) Recherche d’une methode d’estimation despopulations du charancon du bananier, Cosmopolites sordidusGermar (Col., Curculionidae). Acta Oecol.: Oecol. Appl. 1,83–92.

Delattre, P. and Jean-Bart, A. (1978) Activites des champignonsentomopathogenes (Fungi imperfecti) sur les adultes deCosmopolites sordidus Germ. (Coleoptera, Curculionidae).Turrialba 28, 287–93.

de Souza, V., A.F., Warumby, J.F., de Moura, R.J.M.,de Almeida, J.L. and Dantas, A.P. (1981) Dinamica popula-cional de Cosmopolites sordidus (Germar, 1824) e Metamasiushemipterus, e ocorrencia de epizootias por Beauveria bassianaem plantios de bananeira ‘Prata’ situados em topografia devarzea e de serra, no estado de Pernambuco. IPA Divulga 3,252–68.

de Villiers, E.A. (1973) The banana root borer, Cosmopolitessordidus Germar. Banana series No. K.1. South Africa: CSFRI,Nelspruit, 3 pp.

Dochez, C. (1998) Study on Pest Status and Control ofCosmopolites sordidus (Germar) in South Africa, 65 pp.Masters thesis, Heriot-Watt U., Edinburgh, Scotland.

Durans Pinheiro, J.C. and Batista de Carvalho Filho, W. (1985)Flutuacao populacional de Cosmopolites sordidus em bananaisno Maranhao. Comun. Tec. 8, 7.

Edge, V.E. (1974) Cyclodiene-BHC resistance in Cosmopolitessordidus (Germ.) (Coleoptera:Curculionidae) in New SouthWales, Australia. Bull. Entomol. Res. 64, 1–7.

Edge, V.E., Wright, W.E. and Goodyear, G.J. (1975)The development and distribution of dieldrin resistancein banana weevil borer, Cosmopolites sordidus Germar(Coleoptera:Curculionidae) in New South Wales. J. Aust.Entomol. Soc. 14, 165–9.

Edwards, W.H. (1925) La charancon du bananier Cosmopolitessordidus Germar. Rev. Agric. Sucr. Ile Maurice, Mauritius 7–8,513–4 (cited in Schmitt 1993).

Englberger, K. and Toupu, P. (1983) Banana weevil survey inTonga 1983. Tonga German Plant Protection Project. Mss. 9 pp.

Fargues, J. and Luz, C. (2000). Effects of fluctuating mois-ture and temperature regimes on the infection potential ofBeauveria bassiana for Rhodnius prolixus. J. Invertebr. Pathol.75, 202–11.

Ferreira, R.A. (1995) Aspectos do controle biologico deCosmopolites sordidus (Germar 1824) (Coleoptera:Curculionidae atraves de Beauveria bassiana (Balsamo)Vuillemin (Hyphomycetes), 103 pp. Masters thesis. Univ.Federal do Parana, Brazil.

Ferron, P. (1981) Colonization by the fungi Beauveria andMetarhizium. In H.D. Burges (ed) Microbial Control of Pestsand Plant Diseases, 1970–1980, pp. 456–82. New York: Aca-demic Press.

Figueroa, W. (1990) Biocontrol of the banana root borer weevil,Cosmopolites sordidus (Germar) with Steinermatid nematodes.J. Agric. Univ. Puerto Rico 74, 15–9.

Firman, I.D. (1970). Crop protection problems of banana in Fiji.PANS 16, 625–31.

Fogain, R. and Price, N.S. (1994) Varietal screening of some Musacultivars for susceptibility to the banana weevil, Cosmopolitessordidus (Coleoptera: Curculionidae). Fruits 49, 247–51.

Foreman, P. (1976) Investigation into the resistance of the bananaweevil borer (Cosmopolites sordidus Ger) to dieldrin (1971–1972). In Annual Report 1973, pp. 26–7. Jamiaca BananaBoard, Research and Development Department.


Franzmann, B.A. (1976) Banana weevil borer in NorthQueensland. Queensl. Agric. J. 98, 319–21.

Froggatt, J.L. (1924) Banana weevil borer (Cosmopolitessordidus Chev.). Queensl. Agric. J. 21, 369–78.

Froggatt, J.L. (1925) The banana weevil borer (Cosmopolitessordidus). Queensl. J. Agric. 24, 558–93.

Froggatt, J.L. (1928) The banana weevil borer in Java, with noteson other crop pests. Queensl. Agric. J. 6, 530–41.

Gallego, L. (1956) El picudo o taladrador del platano y del abaca,Cosmopolites sordidus (Germar). Rev. Facul. Nac. Agron. 18,65–72.

Gallo, D. (1978) Manual de entomologia agricola. AgronomiaCeres, Sao Paulo, 531 pp (cited in Batista Filho et al. 1991).

Garcia, F., Gomez, J.E. and Belalcazar, S. (1994) Manejo bio-logico y cultura de Cosmopolites sordidus (Germar) en platano.ACORBAT Mem. XI, 385–95.

Geddes, A.M.W. and Iles, M. (1991) The Relative Importance ofCrop Pests in South Asia. Natural Resources Institute BulletinNo. 39. UK, 102 pp.

Gettman, A.D., Mitchell, W.C., Li, P. and Mau, R.F.L. (1992)A hot water treatment for control of the banana root borer,Cosmopolites sordidus (Germar) (Coleoptera:Curculionidae)in banana planting stock. Proc. Hawai. Entomol. Soc. 31,59–63.

Ghesquiere, M.J. (1924) La maladie des bananiers dans leBas-Congo. Bull. Agric. Congo Belge. Brux. 15, 171–5.

Ghesquiere, M.J. (1925) La maladie du bananier au Congo Belge.Bull. Agric. Congo Belge. 3–4, 556–60.

Godonou, I. (1999) The Potential of Beauveria bassiana forthe Management of Cosmopolites sordidus (Germar, 1824)on Plantain (Musa, AAB), 161 pp. Ph.D. thesis, University ofGhana.

Godonou, I., Green, K.R., Oduro, K.A., Lomer, C.J. andAfreh-Nuamah, K. (2000) Field evaluation of selected formula-tions of Beauveria bassiana for the management of the bananaweevil (Cosmopolites sordidus) on plantain (Musa spp., AABgroup). Biocontrol Sci. Technol. 10, 779–88.

Goitia, W. and Cerda, H. (1998) Hormigas y otras insectos aso-ciados a Musaceas y su relacion con Cosmopolites sordidusGermar (Coleoptera:Curculionidae). Agron. Trop. 48, 209–24.

Gold, C.S. (1998a) Banana weevil: Ecology, pest status andprospects for integrated control with emphasis on East Africa.In R.K. Saini (ed) Proc. Third Int. Conf. Trop. Entomol.,30 October–4 November 1994. Nairobi, pp. 47–71. Nairobi:ICIPE Science Press.

Gold, C.S. (1998b) Integrated pest management of banana weevilwith emphasis on East Africa. In F. Rosales, S.C. Triponand J. Cerna (eds) Proc. Int. Workshop Org. Environ.Friend. Banana Prod. Proc. Workshop Int. Network Improv.Banana Plantain, Guacimo, Costa Rica, July 27–29, 1998,pp. 145–163. Montpellier, France: INIBAP.

Gold, C.S. and Bagabe, M.I. (1997) Banana weevil, Cosmopolitessordidus Germar (Coleoptera, Curculionidae), infestation ofcooking and beer bananas in adjacent stands in Uganda. Afr.Entomol. 5, 103–8.

Gold, C.S. and Gemmill, B. (eds) (1993). Biological andIntegrated Control of Highland Banana and Plantain Pests andDiseases. Proceedings of a Research Coordination Meeting.455 pp. Cotonou, Benin: IITA .

Gold, C.S., Ogenga-Latigo, M.W., Tushemereirwe, W.,Kashaija, I. and Nankinga, C. (1993) Farmer perceptions of

banana pest constraints in Uganda: Results from a rapid ruralappraisal. In C.S. Gold and B. Gemmill (eds) Biological andIntegrated Control of Highland Banana and Plantain Pests andDiseases. Proceedings of a Research Coordination Meeting,pp. 3–24. Cotonou, Benin: IITA.

Gold, C.S., Speijer, P.R., Karamura, E.B. and Rukazambuga, N.D.(1994a) Assessment of banana weevils in East African highlandbanana systems and strategies for control. In R.V. Valmayor,R.G. Davide, J.M. Stanton, N.L. Treverrow and V.N. Roa(eds) Proceedings of Banana Nematode/Borer Weevil Conf.Kuala Lumpur, 18–22 April 1994, pp. 170–90. Los Banos,Philippines.

Gold, C.S., Speijer, P.R., Karamura, E.B., Tushemereirwe, W.K.and Kashaija, I.N. (1994b) Survey methodologies for pest anddisease assessment in Uganda. Afr. Crop Sci. J. 2, 309–21.

Gold, C.S., Okech, S.H. and Ssendege, R. (1997) Banana weevilpopulation densities and related damage in Ntungamo andMbarara districts, Uganda. In E. Adipala, J.S. Tenywa andM.W. Ogenga-Latigo (eds) Afr. Crop Sci. Conf. Proc.. Pretoria,13–17 January 1997, pp. 1207–19. Makerere University,Kampala, Uganda.

Gold, C.S., Night, G., Abera, A. and Speijer, P.R. (1998a) Hot-water treatment for control of banana weevil, Cosmopolitessordidus Germar (Coleoptera:Curculionidae) in Uganda. Afr.Entomol. 6, 215–21.

Gold, C.S., Night, G., Speijer, P.R., Abera, A.M.K. andRukazambuga, N.D.T.M. (1998b) Infestation levels of bananaweevil, Cosmopolites sordidus Germar, in banana plants estab-lished from treated propagules in Uganda. Afr. Entomol. 6,253–63.

Gold, C.S., Bagabe, M.I. and Ssendege, R. (1999a). Bananaweevil, Cosmopolites sordidus (Germar): (Coleoptera:Curculionidae) tests for suspected resistance to carbofuran anddieldrin in Masaka District, Uganda. Afr. Entomol. 7, 189–96.

Gold, C.S., Karamura, E.B., Kiggundu, A., Bagamba, F. andAbera, A.M.K. (1999b) Geographic shifts in highland cookingbanana (Musa spp., group AAA-EA) production in Uganda.Int. J. Sustain. Dev. World Ecol. 6, 45–59.

Gold, C.S., Nemeye, P. and Coe, R. (1999c) Recognition and dura-tion of larval instars of banana weevil, Cosmopolites sordidusGermar, in Uganda. Afr. Entomol. 7, 49–62.

Gold, C.S., Rukazambuga, N.D.T.R., Karamura, E.B., Nemeye, P.and Night, G. (1999d) Recent advances in banana weevil biol-ogy, population dynamics and pest status with emphasis on EastAfrica. In: E. Frison, C.S. Gold, E.B. Karamura and R.A. Sikora(eds) Mobilizing IPM for Sustainable Banana Production inAfrica. Proceedings of a Workshop on Banana IPM, Nelspruit,South Africa, 23–28 November 1998, pp. 33–50. Montpellier,France: INIBAP.

Gold, C.S., Kagezi, G., Nemeye, P. and Ragama, P. (2002a)Density effects of the banana weevil Cosmopolites sordidus(Germar) on its oviposition performance and egg and larvalsurvivorship. Insect Sci. Appl. 22, 205–13.

Gold, C.S., Okech, S.H. and S. Nokoe (2002b) Evaluation of pseu-dostem trapping as a control of banana weevil, Cosmopolitessordidus (Germar), populations and damage in Ntungamodistrict, Uganda. Bull. Entomol. Res. 92, 35–44.

Gomes, C. (1985) Estudo do comportamento da brocada bananeira Cosmopolites sordidus (German, 1824)(Coleoptera: Curculionidae), visando seu controle, 82 pp.Doctor in Sciences thesis. U. do Sao Paulo, Brazil.


Gordon, J. and Ordish, G. (1966) Insect pests of banana. InPANS Manual No. 1. Bananas, pp. 33–5. London: Ministryof Overseas Development.

Gorenz. A.M. (1963) Preparation of disease free plantingmaterials of banana and plantain. Ghan. Farm. 7, 15–8.

Gowen, S.R. (1995) Pests. In S. Gowen (ed) Bananas andPlantains, pp. 382–402. London: Chapman and Hall.

Gravier, C. (1907) Sur un coleoptore (Sphenoporous striatusFahr) qui attaque les bananiers a Sau Thome (Golfe de Guinee).Bull. Mus. Nat. d’Hist, Paris 13, 30–2 (cited in Viswanath1976).

Greathead, D.J. (1986) Opportunities for biological control ofinsect pests in tropical Africa. Revue. Zool. Afr. 100, 85–96.

Greathead, D.J., Cock, M.J.W. and Girling, D.J. (1986) DraftReport on a Consultancy for IITA to Assess the Potential forBiological Control of Pests of the African Food Crops: Maize,Sorghum, Rice, Plantain, Cowpea, Sweet Potato and Cassava.Ibadan, Nigeria: IITA, 74 pp.

Gressitt, J.L. (1954) Insects of Micronesia. Vol. 1. Honolulu:Bishop Museum.

Griesbach, M. (1999). Occurrence of Mutualistic Endophytesin Bananas (Musa spp.) and their Potential as BiocontrolAgents of the Banana Weevil Cosmopolites sordidus (Germar)(Coleoptera: Curculionidae) in Uganda, 131 pp. Ph.D. thesis,University of Bonn.

Haarer, A.A. (1964) Modern Banana Production. London:Leonard Hill, 134 pp.

Haddad, O., Surga, J. and Wagner, M. (1979) Relacion de lacomposicion genomica de las musaceas con el grado de atrac-cion de adultos y danos de larvas de Cosmopolites sordidus G.(Coleoptera:Curculionidae). Agron. Trop. 29, 429–38.

Hall, W.J. (1954) Insect pests in British colonial dependencies:A half yearly report. FAO Plant Prot. Bull. 2, 81–2.

Hamill, R.L., C.E. Higgins, H.E. Boaz and M. Gorman (1969)The structure of beauvericin, a new depsipeptide antibiotic toArtemia salina. Tetrahedron Lett. 49, 4255–8.

Hargreaves, H. (1940) Insect pests of bananas. In J.D. Tothill(ed) Agriculture in Uganda, pp. 121–4. Oxford, UK: OxfordUniversity Press.

Harris, W.V. (1947) The banana borer. East Afr. Agric. For. J. 13,15–8.

Hassan, E. (1977) Major Insect and Mite Pest of Australian crops.Gatton, Queensland: Ento. Press.

Hasyim, A. and Gold, C.S. (1999) Potential of classical biologi-cal control for banana weevil, Cosmopolites sordidus Germar,with natural enemies from Asia (with emphasis on Indonesia).In E. Frison, C.S. Gold, E.B. Karamura and R.A. Sikora(eds) Mobilizing IPM for Sustainable Banana Production inAfrica. Proceedings of a Workshop on Banana IPM, Nelspruit,South Africa, 23–28 November 1998, pp. 59–71. Montpellier:INIBAP.

Hely, P.C., Pasfeld, G. and Gellatley, J.G. (1982) Insect Pestsof Fruit and Vegetables in New South Wales. Melbourne,Australia: Inkata Press.

Hildreth, R.C. (1962) Certified banana seed. Trop. Agric. 39,103–5.

Hord, H.H.V. and Flippin, S.J. (1956) Studies of banana weevilsof Honduras. Econ. Entomol. 49, 296–300.

Hoyt, C.P. (1957) Parasites and Predators Introduced into thePacific Islands for the Biological Control of Insects and OtherPests. South Pacific Commission Technical Paper No. 101,40 pp.

ICIPE (1991) Annual report: 1990. Nairobi, Kenya: ICIPE.Ingles, R. and Rodriguez, J. (1989) Evaluacion de plagaci-

das y metodos para combatir el picudo negro del platano(Cosmopolites sordidus Germar). J. Agric. Univ. Puerto Rico73, 97–107.

INIBAP (1988a) Plantain in Western Africa. INIBAP internaldocument 88/1. Montpellier, France.

INIBAP (1988b.) Nematodes and the Borer Weevil in Bananas:Proc. Workshop, 7–11 December 1987, Bujumbura, Burundi.

Irizzary, H., Rivera, E., Rodriguez, J., Beauchamp de Caloni, I.and Oramas, D. (1988) The Lacknau plantain: A high yieldingcultivar with field resistance to the corm weevil, Cosmopolitessordidus (Germar). J. Agric. Univ. Puerto Rico 72, 353–63.

Ittyeipe, K. (1986) Studies on the host preference of bananaweevil borer Cosmopolites sordidus GERM. (Curculionidae –Coleoptera). Fruits 41, 375–9.

Jaramillo, R. (1979) Algunos aspectos agronoBujumbura,Burundi,micos del cultivo del banana y del platano InCATIE-UC/USAID-OIRSA Control Integrado de Plagasen Sistemas de Produccion de Cultivos para PequenosAgricultores, pp. 271–95. Turrialba, Costa Rica.

Jardine, N.K. (1924) Plantain root beetle borer (Cosmopolitessordidus Germar). Trop. Agric. 62, 6.

Jayaraman, S., Ndiege, I.O., Oehlschlager, A.C., Gonzalez, L.M.,Alpizar, D., Falles, M., Budenberg, W.J. and Ahuya, P.(1997) Synthesis, analysis, and field activity of sordidin, amale-produced aggregation pheromone of the banana weevil,Cosmopolites sordidus. J. Chem. Ecol. 23, 1145–61.

Jepson, F.P. (1914) A Mission to Java in Quest of Natural Enemiesfor a Coleopterous Pest of Bananas (Cosmopolites sordidus,Chevr.). Fiji Department of Agriculture. Bulletin No. 7, 23 pp.

Jirasuat, M. et al. (1989) Study of biology of banana cormborer weevil. Annual report of the Entomology and ZoologyDivision. Bangkok, Thailand: Department of Agriculture, 5 pp.(in Thai) (cited by Vittayaruk et al. 1994).

Job, S.C., Yagean, T., Venkitesan, T.S. and Abraham, C.C. (1986)Integrated control of the rhizome weevil and burrowing nema-tode infesting banana var. Nendran. Pesticides 20(10), 10–1.

Jones, D.E. (1968) Attraction of banana rhizome volatiles tothe banana root borer, Cosmopolites sordidus Germar. Unpub-lished manuscript. United Fruit Company, La Lima, Honduras,7 pp.

Jones, M.T. (1986) Pests and diseases of bananas and plantains ofTrinidad and Tobago. J. Agric. Soc. Trinidad and Tobago 86,18–33.

Jurado, R. (1974) Manual practico del cultivo del bananocavendish. Uraba, Anitoquia. Mimeo (cited by Arroyave 1985).

Kaaya, G.P., Seshu Reddy, K.V., Kokwaro, E.D. andMunyinyi, D.M. (1993) Pathogenicity of Beauveria bassiana,Metarhizium anisoplae and Serratia marcescens to the bananaweevil Cosmopolites sordidus. Biocontrol Sci. Technol. 3,177–87.

Karamura, D.A. (1998) Numerical Taxonomic Studies of theEast African Highland Bananas (Musa AAA-East Africa) inUganda, 344 pp. Ph.D. thesis. University of Reading, UK.

Kehe, M. (1985) Les principaux insectes depredateurs du plan-tain en Cote D’Ivoire: Importance des infestations et incidenceagro-economique. In Int. Assoc. Res. Plantain Banana Meet.,27–31 May 1985, pp. 94–101. Abidjan, Cote d’Ivoire.

Kehe, M. (1988) Le charancon du bananier (Cosmopolitessordidus) les acquis et les perspectives de la recherche:


Contribution de l’IRFA-CIRAD/Cote d’Ivoire. In Nematodesand the Borer Weevil in Bananas: Proceedings of a Workshop,7–11 December 1987, Bujumbura, Burundi, pp. 47–53.Montpellier: INIBAP.

Kelly D.S. (1966) Control of dieldrin resistant banana weevilborer. Unpublished manuscript. Queensland Dept. of MiningIndustries, Australia, 7 pp.

Kermarrec, A. and Mauleon, H. (1975) Controle biologiqueexperimental de Cosmopolites sordidus par la RhabditideNeaplectana carpocapsae (Nematoda: Neoaplectanidae).Proc. 8th OTAN Congr., 4–5 August 1975, St. Lucia (cited inKermarrec et al. 1993).

Kermarrec, A. and Mauleon, H. (1989) Synergie entre lechlordecone et Neoaplectana carpocapsae Weiser (Nematoda:Steinermatidae) pour le controle de Cosmopolites sordidus(Coleoptera: Curculionidae). Rev. Nematol. 12, 324–5.

Kermarrec, A., Sirjusingh, C., Mauleon, H., Pavis, C. andSarah, J.L. (1993) Biological control of weevils and white-grubs in the Caribbean: A review. In C.S. Gold and B. Gemmill(eds) Biological and Integrated Control of Highland Bananaand Plantain Pests and Diseases. Proceedings of a ResearchCoordination Meeting, pp. 155–70. Cotonou, Benin: IITA.

Khan, A. and Gangapersad, G. (2001) Comparison of the effec-tiveness of three entomopathogenic fungi in the manage-ment of banana borer weevil, Cosmopolites sordidus (Germar)(Coleoptera: Curculionidae). Int. Pest Control 43, 208–13.

Kiggundu, A. (2000) Host Plant Reactions and ResistanceMechanisms to Banana Weevil, Cosmopolites sordidus(Germar) in Ugandan Musa germplasm, 98 pp. Masters thesis.Orange Free State University, South Africa.

Kiggundu, A., Vuylsteke, D. and Gold, C.S. (1999) Recentadvances in host plant resistance to banana weevil,Cosmopolites sordidus Germar. In E. Frison, C.S. Gold,E.B. Karamura and R.A. Sikora (eds) Mobilizing IPM forSustainable Banana Production in Africa. Proceedings of aWokshop on Banana IPM, 23–28 November 1998, Nelspruit,South Africa, pp. 87–96. Montpellier, France: INIBAP.

Knowles, L.H. and Jepson, F.P. (1912) Department of Agriculture.Fiji Bulletin 17 pp. (cited by Cuille 1950).

Koppenhofer, A.M. (1993a) Observations on egg-layingbehaviour of the banana weevil, Cosmopolites sordidus(Germar). Entomologia Experimenalis et Applicata 68, 187–92.

Koppenhofer, A.M. (1993b) Egg predators of the banana weevil,Cosmopolites sordidus (Germar) (Col., Curculionidae) inWestern Kenya. Journal of Applied Entomology 116, 352–57.

Koppenhofer, A.M. (1993c) Search and evaluation of natural ene-mies of the banana weevil. In: C.S. Gold and B. Gemmill (eds).Biological and Integrated Control of Highland Banana andPlantain Pests and Diseases. Proceedings of a Research Coor-dination Meeting, pp 87–96. IITA. Contonou, Benin.

Koppenhofer, A.M. (1994) Observations on the bionomics ofThyreocephalus interocularis (Eppelsheim) (Col., Staphyl-inidae), a predator of the banana weevil. Journal of AppliedEntomology 117, 382–94.

Koppenhofer, A.M. (1995) Bionomics of Euborellia annulipes inWestern Kenya (Dermaptera: Carcinophoridae) EntomologiaGeneralis 20, 81–6.

Koppenhofer, A.M. and Schmutterer, H. (1993) Dactylosternumabdominale (F.) (Coleoptera: Hydrophilidae): A predator of thebanana weevil. Biocontrol Science and Technology 3, 141–7.

Koppenhofer, A.M., Seshu Reddy, K.V. (1994) A comparison ofrearing methods for the banana weevil, Cosmopolites sordidus(Germar) (Coleoptera: Curculionidae) on its natural host. InsectScience and its Application 15, 191–5.

Koppenhofer, A.M., Seshu Reddy, K.V., Madel, G. andLubega, M.C. (1992) Predators of the banana weevil,Cosmopolites sordidus (Germar) (Col., Curculionidae) inWestern Kenya. J. Appl. Entomol. 114, 530–3.

Koppenhofer, A.M., Seshu Reddy, K.V. and Sikora, R.A. (1994)Reduction of banana weevil populations with pseudostem traps.Inter. J. Pest Manage. 4, 300–4.

Koppenhofer, A.M., Sikora, R.A. and Seshu Reddy, K.V.(1995) Eidonomy and ecology of Dactylosternum abdominale(Coleoptera: Hydrophilidae), a predator of banana weevilCosmopolites sordidus (Coleoptera: Curculionidae). Entomol.Gen. 19, 303–13.

Kusomo, S. and Sunaryono, H. (1985) Status of banana pro-duction in Indonesia. In B.E. Umali and C.M. Lancitan (eds)Banana and Plantain Research and Development, pp. 35–8.Los Banos, Philippines: ACIAR and PCARRD.

Laumond, C., Mauleon, H. and Kermarrec, A. (1979) Donnesnouvelle sur le spectre d’hotes et le parasitisme du nema-tode entomphage Neoaplectana carpocapsae. Entomophaga24, 13–27.

Lemaire, L. (1996) Les relations semiochimiques chez lecharancon Cosmopolites sordidus (Germar)(Coleoptera: Curculionidae) et la resitance de sa plante-hote,le bananier., 268 pp. Ph.D. thesis. University of Montpellier,France.

Lescot, T. (1988) Influence de l’altitude sur les populationsdu charancon des bananiers (Cosmopolites sordidus Germar).Fruits 43, 433–7.

Liceras, L., Urrelo, G. and Beltran, F. (1973) Ensayo para elcontrol del gorgojo negro del platano, Cosmopolites sordidusGermar (Coleoptera:Curculionidae), al momento de la siembra.Rev. Peru. de Entomol. 16, 50–4.

Lino Neto, J. and Dolder, H. (1995) Characteristics ofthe spermatazoon of Cosmopolites sordidus (Coleoptera:Curculionidae). In B.G.M. Jamieson, J. Ausio and J.-L. Justine(eds) Advances in Spermatozoal Phylogeny and Taxonomy.Mem. Mus. Natn. Hist. Nat., pp 297–300, Vol. 166. Paris.

Litsinger, J.A. (1974) Final Report of the Entomologist,1972–1974. Ministry of Agriculture, Fisheries and Forestry,Tonga, 63 pp.

Loebel, R. (1975) Weevil borer not main cause of plantationdecline. Banana Bull. 39(7), 10.

Londono, M.E., Pulido, J.I., Garcia, F., de Ploania, I.Z. andLeon, G. (1991) Manejo integrado de plagas. In: S.L. Belacazar(ed.) El Cultivo del Platano (Musa ABB Simmonds) en elTropico, pp. 310–26. Colombia: ICA.

Longoria, A. (1968) Diferencias sexuales en la morfolo-gia externa de Cosmopolites sordidus Germar (Coleoptera,Curculionidae). Cienc. Biol. Hab. 1, 1–11.

Longoria, A.G.G. (1972) Crianza en el laboratorio y datos preli-marios sobre el ciclo de vida de Cosmopolites sordidus Germar(Col. Curculionidae). Nota de las Ciencias Biologicas No. 30.In Primer Seminario de Investigaciones de la Facultad deCiencias. Centro de Informacion Cientifica y Tecnica, Uni-versidad de la Habana, Cuba.

Marcelino, L. and Quintero, J.A. (1991) Influencia del picudonegro (Cosmopolites sordidus) y de la precipitacion en los


platanares de cuatro localidades de Baru, Chiriqui. Rev. Cienc.Agropec. 7, 49–57.

Martinez, J.A. (1971) Flutuacoes da populacao da broca-da-bananeira ‘Moleque’ (Cosmopolites sordidus Germar). Anaisdo I Congreso Bras. de Fruticultura, pp. 187–94.

Martinez, M. and Longoria, A. (1990) Oviposicion y desarrollo deCosmopolites sordidus (Coleoptera: Curculionidae) en platanoy malanga. Cienc. Agric. 40, 169–71.

Masanza, M. (1995) Integrating Pseudostem Trapping, Chemi-cal and Biological Control for the Management of the BananaWeevil (Cosmopolites sordidus Germar), 93 pp. Masters thesis.Makerere University, Kampala, Uganda.

Masanza, M. (1999) End of Year Progress Report on Ph.D.Thesis Proposal, Effect of Crop Residue Management Prac-tices on Banana Weevil (Cosmopolites sordidus) Populationsand Associated Damage, 43 pp. Kampala, Uganda: IITA.

Masso, E. and Neyra, M. (1997) Danos y perdidas causadas porCosmopolites sordidus en el cultivo del platano. Agrotec. Cuba27, 86–8.

Mau, R.F.L. (1981) The banana root borer, a new pest. HawaiiCooperative Extension Service Entomological Notes No. 11.4 pp.

Mbwana, A.S.S. and Rukazambuga, N.D.T.M. (1999) BananaIPM in Tanzania. In E. Frison, C.S. Gold, E.B. Karamuraand R.A. Sikora (eds) Mobilizing IPM for SustainableBanana Production in Africa. Proceedings of a Workshop onBanana IPM, Nelspruit, South Africa, 23–28 November 1998,pp. 237–45. Montpellier, France: INIBAP.

McCarthy, T. (1920) Banana root borer (Cosmopolites sordidusGermar) Agric. Gaz. NSW 31, 865–72. (cited in Viswanath1976).

McIntyre, B.D., Gold, C.S., Kashaija, I.N., Ssali, H., Night, G.and Bwamiki, D.P. (2002) Effects of legume intercrops on soil-borne pests, biomass, nutrients and soil water in banana. Biol.Fertility Soils 34, 342–8.

McNutt, D. (1974) A review of banana weevil control in Uganda,with further tests of dieldrin formulations. East Afr. Agric. For.J. 39, 205–9.

Medina, G., Garcia, T. and Martorell, L. (1975) Preliminaryscreening of pesticides for control of banana roots borer,Cosmopolites sordidus Germar (Coleoptera: Curculionidae).J. Agric. Univ. PR 59, 79–81.

Mello, E.J.R., Mello, R.H. and Sampaio, A.S. (1979) Resistenciaao aldrin em brocas de bananeira Cosmopolites sordidus Germ.do litoral paulista. Biologico (Brasil) 45, 249–54.

Mesquita, A.L.M. (1985) Avaliacao do ataque do Cosmopolitessordidus (Germar, 1824) (Col.: Curculionidae) em rizoma debananeira. Pesq. Andam. 21, 2.

Mesquita, A.L.M. (1988) Controle biologico das brocasda bananeira Cosmopolites sordidus (Germar, 1824) eMetamasius hemipterus (Linne, 1764) com fungos ento-mogenos. In Santa Marta, Colombia 1987, Mem. ACORBATVIII, pp. 311–24.

Mesquita, A.L.M. and Alves, E.J. (1983) Aspectos da biolo-gia da broca-do-rizoma em diferentes cultivares de bananeira(Cosmopolites sordidus, Musa acuminata). Pesq. Agropec. Bra.18, 1289–92.

Mesquita, A.L.M. and Alves, E.J. (1984) Inimigos naturais deCosmopolites sordidus e Metamasius hemipterus no Brasil.Rev. Brasil. Fruitic. (Cruz das Almas) 6, 45–6.

Mesquita, A.L.M. and Caldas, R.C. (1986) Efeito da idadee da cultivar de bananeira sobre a biologia e preferen-cia do Cosmopolites sordidus (Germar, 1824) (Coleoptera,Curculionidae). Fruits 41, 245–9.

Mesquita, A.L.M., Lucchini, F., Alves, E.J. and Caldas, R.C.(1981) Influencia dos fatores ambientais no grau de para-sitismo de Beauveria bassiana sobre Cosmopolites sordidus eMetamasius hemipterus, em cultivo da bananeira. Pesq. Andam.14, 4.

Mesquita, A.L.M., Alves, E.J. and Caldas, R.C. (1984) Resistanceof banana cultivars to Cosmopolites sordidus (Germar 1824).Fruits 39, 254–7.

Messiaen, S. (2000) Neem (Azadirachta indica), wood ashes,offee husk and hot pepper (Capsicum spp.) for controllingthe banana weevil (Cosmopolites sordidus): Investigations intotheir effect and mode of action. Unpublished manuscript.Centre de Recherches Regionales sur Bananiers et Plantains.Njombe, Cameroon. 13 pp.

Messiaen, S. (2002). Components of a strategy for the integratedmanagement of the banana weevil Cosmopolites sordidus(Germar) (Coleoptera: Curculionidae). Dissertationes de Agri-culture No. 540. PhD thesis, Faculty of Agricultural andApplied Biological Sciences, Catholic University Leuven,Belgium 169 pages.

Messiaen, S., Fogain, R., Ysenbrandt, H. and Sama Lang, P.(2000) In situ efficacy of neem (Azadirachta indica A. Juss)for controlling banana weevil Cosmopolites sordidus Germar(Coleoptera: Curculionidae). Unpublished manuscript. Centrede Recherches Regionales sur Bananiers et Plantains. Njombe,Cameroon. 13 pp.

Mestre, J. (1995) Reconnaissance de sexes chez le charancon dubananier Cosmopolites sordidus (Germar, 1824) (ColeopteraCurculionidae). CIRAD-FLHOR, Station de Neufchateau.Note Technique 1. 8 pp.

Mestre, J. (1997) Les recherches recentes sur le charancon desbananiers, Cosmopolites sordidus (Germar, 1824) (Coleoptera:Curculionidae). Fruits 52, 67–82.

Mestre, J. and Rhino, B. (1997) Les etudes sur le charancondes banaiers, Cosmopolites sordidus (Germar, 1924): Bilansommaire – Neufchateau 1995–1997. CIRAD-FLHOR docu-ment JM-97-04. Neufchateau. 20 pp. + annexes.

Minost, C. (1992) Etude de la communication semiochemiquechez le charancon du banaier, Cosmopolites sordidus Germar(1824) (Coleoptera: Curculionidae). DAA thesis, InstitutNational Agronomique, Paris.

Mitchell, G. (1978) The Estimation of Banana Borer Populationand Resistance Levels. Technical Bulletin 2, Windward IslandBanana Growers Association (WINBAN), St Lucia. 34 pp.

Mitchell, G. (1980) Banana Entomology in the Windward Islands.Final Report 1974–1978. Windward Island Banana GrowersAssociation (WINBAN), St Lucia, 216 pp.

Montellano, B. (1954) Estudios biologicos del Cosmopolitessordidus Germar que infesta al rizoma de abaca. 27 pp. TesisMag. Agr. Turrialba, Costa Rica: IICA.

Moreira, R.S. (1971) A broca das banaeiras. Corr. Agric.(Sao Paulo) 1, 10–12.

Moreira, R.S. (1979) Bananis livres de broca produzem o dobro.Corr. Agric. 2, 202–6.

Moreira, R.S., Laurencao, A.L. and Saes, L.A. (1986) Compara-cao entre o queijo e a telha como iscas na atratividade domoleque das bananeiras. In Congr. Bras. Fruticult., 8, Brasilia,26–31 enero. SBF, pp. 87–92.


Mori, K., Kakayama, T. and Takikawa, H. (1996) Synthesis andabsolute configuration of sordidin, the male-produced aggrega-tion pheromone of the banana weevil Cosmopolites sordidus.Tetrahedron Lett. 37, 3741–4

Moznette, G.F. (1920) Banana root-borer. J. Agric. Res. 19,39–46.

Mukandala, L.G., Ndile, M.R., Sentozi, E. and Bosch, C.H.(1994) Planning of Participatory Research in Ntoija, BukobaDistrict, Tanzania. Tanzania/Netherlands Farmings SystemsResearch Project Lake Zone. Field Note No. 46. AgriculturalResearch Institute, Maruku, Tanzania, 33 pp.

Musabyimana, T. (1995) Studies on the Banana Weevil(Cosmopolites sordidus) and Nematode Complex in WesternKenya: March–October 1995. Final Report. ICIPE, Nairobi,Kenya, 23 pp.

Musabyimana, T. (1999) Neem Seed for the Management of theBanana Weevil, Cosmopolites sordidus Germar (Coleoptera:Curculionidae) and Banana Parasitic Nematode Complex, 175pp. Ph.D. thesis. Kenyatta University, Nairobi, Kenya.

Musabyimana, T., Saxena, R.C., Kairu, E.W., Ogol, C.P.K.O. andKhan, Z.R. (2001) Effects of neem seed derivatives on behav-ioral and physiological responses of the Cosmopolites sordidus(Coleoptera: Curculionidae). Hortic. Ent. 94, 449–54.

Nahif, A.A. (1998) Morphology, histology and histochemistry ofthe reproductive system of Cosmopolites sordidus (Coleoptera:Curculionidae). Part 2. The female internal genitalia. Unpub-lished manuscript. University of Bonn, 13 pp.

Nahif, A.A. (2000). An anatomical, histological, and his-tochemical study of the male reproductive organs ofCosmopolites sordidus (Germar) (Coleoptera: Curculionidae).Beit.-Entomol. 50, 271–81.

Nahif, A.A., Koppenhofer, A. and Madel, G. (1994) Morpholo-gie, biologie und bedeuntung von Cosmopolites sordidus,Germar, 1824 (Coleoptera: Curculionidae. Zeits. Angwe. Zool.4, 435–47.

Nankinga, C.M. (1994) Potential of Indigenous Fungal Pathogensfor the Biological Control of the Banana Weevil, Cosmopolitessordidus (Germar), in Uganda, 95 pp. Masters thesis, MakerereUniversity, Kampala, Uganda.

Nankinga, C.M. (1997) Characterisation of Beauveria andMetarhizium Isolates and the Influence of Delivery Sys-tems on their Use as Biological Control Agents of BananaWeevil, Cosmopolites sordidus. University of Reading upgrad-ing report. Reading, UK.

Nankinga, C.M. (1999) Characterization of EntomopathogenicFungi and Evaluation of Delivery Systems for the BiologicalControl of the Banana Weevil, Cosmopolites sordidus, 277 pp.Ph.D. thesis, University of Reading, UK.

Nankinga, C.M. and Ogenga-Latigo, M.W. (1996) Effect ofmethod of application on the effectiveness of Beauveriabassiana against the banana weevil, Cosmopolites sordidus.Afr. J. Plant Prot. 6, 12–21.

Nankinga, C., Bridge, P., Karamura, E. and Moore, D. (1996)Biochemical characterisation of fungal pathogens isolated frombanana fields in Uganda and their prospects for biologicalcontrol of the banana weevil, Cosmopolites sordidus. Unpub-lished mss. Kawanda Agricultural Research Institute, Kampala,Uganda.

Nankinga, C.M., Moore, D., Bridge, P. and Gowen, S. (1999)Recent advances in microbial control of banana weevil. In E.

Frison, C.S. Gold, E.B. Karamura and R.A. Sikora (eds) Mobi-lizing IPM for Sustainable Banana Production in Africa. Pro-ceedings of a Workshop on Banana IPM, Nelspruit, SouthAfrica, 23–28 November 1998, pp. 73–85. Montpellier, France:INIBAP.

Nanne, H.W. and Klink, J.W. (1975) Reducing banana root weeviladults from an established banana plantation. Turrialba 25,177–9.

Nanthachai, P. (1985) Banana production and research programsin Thailand. In B.E. Umali and C.M. Lancitan (eds) Banana andPlantain Research and Development, pp. 45–51. Los Banos,Philippines: ACIAR and PCARRD.

Ndiege, I.O., Budenberg, W.J., Lwande, W. and Hassanali, A.(1991) Volatile components of banana pseudostem of a cultivarsusceptible to the banana weevil. Phytochemistry 30, 3929–30.

Ndiege, I.O., Budenberg, W.J., Otieno, D.O. and Hassanali, A.(1996a) 1,8-Cineole: An attractant for the banana weevil,Cosmopolites sordidus. Phytochemistry 42, 369–71.

Ndiege, I.O., Jayaraman, S. and Oehlschlager, A.C. (1996b) Con-venient synthesis and field activity oa a male-produced aggre-gation pheromone of Cosmopolites sordidus. Phytochemistry42, 280–2.

Ndege, L.J., Ndile, M.R. and Bosch, C.H. (1995) Farmers’Assessment of a Weevil Trapping Trial in Ntoija Village, BukobaDistrict. Lake Zone Farming Systems project, AgriculturalResearch Institute, Maruku, Tanzania. Progress Report No. 9.10 pp.

Neuenschwander, P. (1988) Prospects and proposals for biolog-ical control of Cosmopolites sordidus (Germar) (Coleoptera:Curculionidae) in Africa. In Nematodes and the Borer Weevilin Bananas: Proceedings of a Workshop, Bujumbura, Burundi,7–11 December 1987, pp. 54–57. Montpellier, France: INIBAP.

Ngode, L. (1998) Management of Banana Weevil CosmopolitesSordidus Germar (Coleoptera: Curculionidae) for ImprovedBanana Yield in Western Kenya, 180 pp. Ph.D. thesis, KenyattaUniversity, Nairobi, Kenya.

Niere, B.I. (2001). Significance of Non-Pathogenic Isolates ofFusarium oxysporum Schlecht:Fries for the Biological Controlof the Burrowing Nematode Radopholus similis (Cobb) Thorneon Tissue Cultured Banana, 118 pp. + annexes. Ph.D. thesis,University of Bonn.

Nkakwa, A.A. (1999) Susceptibility of some plantain cultivarsthe plantain/banana weevil, Cosmopolites sordidus Germar(Coleoptera: Curculionidae), 73 pp.+annexes. Masters thesis,University of Ghana, Legon.

Nonveiller, G. (1965) Comment proteger les bananiers contre lesattaques du charancon. Camer. Agric. Past. For. 87, 32–43.

Ochieng, V.O. (2001) Genetic Biodiversity In banana WeevilCosmopolites Sordidus Populations in Banana GrowingRegions of the World, 139 pp. Ph.D. thesis, University ofNairobi.

Ogenga-Latigo, M.W. (ed.) (1992) Recent Contribution toBanana Entomology in Uganda (1990–1992), 52 pp. Depart-ment of Crop Sciences, Makerere University, Kampala,Uganda.

Ogenga-Latigo, M.W. and Bakyalire, R. (1993) Use of pseu-dostem traps and coefficient of infestation (PCI) for assess-ing banana infestation and damage by Cosmopolites sordidusGermar. Afr. Crop Sci. J. 1, 39–48.

Okech, S.O., Gold, C.S., Karamura, E.B., Ssali, H. and Speijer, P.(1996) Banana Weevil and Nematode IPM Project: First


Annual Report (August 1995–July 1996) 40 pp. African High-lands Initiative. Nairobi, Kenya: ICRAF.

Oliveira, A.M. de, Sudo, S., Barcellos, D.F., Mendes, S.G.,Maiolino, W. and do A. Meneguelli, N. (1976) Fluctuacao dapopulacao de Cosmopolites sordidus e Metamasius spp. embananais de Agra dos Reis, Estado do Rio de Janeiro. Pesq.Agropec. Bras., Ser. Agron. 11, 37–41.

Ortiz, R., Vuylsteke, D., Dumpe, B. and Ferris, R.S.B. (1995)Banana weevil resistance and corm hardness in Musagermplasm. Euphytica 86, 95–102.

Ostmark, H.E. (1974) Economic insect pests of bananas. Annu.Rev. Entomol. 19, 161–76.

Padmanaban, B., Sundararaju, P., Velayudhan K.C. andSathiamoorthy S. (2001) Evaluation of Musa germplasmagainst banana weevil borers. Infomusa 10, 26–8.

Painter, R.H. (1951) Insect Resistance in Crop Plants, 512 pp.New York: The MacMillan Co.

PANS (1973) Pest Control in Bananas. Pans Manual No. 1. ThirdEdition. London: Ministry of Overseas Development, 126 pp.

Parnitzki, P. (1992) Biologische Bekampfund des Russelka-fers Cosmopolites sordidus (Germar) mit entomopathogenenNematoden der Gattungen Heterorhabditis und Steinernemasowie Untersuchungen zur Biologie des Schadlings, 131 pp.Ph.D. thesis, University of Bonn.

Pavis, C. (1988) Quelques aspects comportementaux chezle charancon du bananier Cosmopolites sordidus Germar(Coleoptera: Curulionidae). In Nematodes and the Borer Weevilin Bananas: Proceedings of a Workshop, Bujumbura, Burundi,7–11 December 1987, pp. 58–61. Montpellier, France: INIBAP.

Pavis, C. (1993) Etude des relations plante-insecte chez le cha-rancon du bananier Cosmopolites sordidus. In C.S. Gold and B.Gemmill (eds) Biological and Integrated Control of HighlandBanana and Plantain Pests and Diseases. Proceedings of aResearch Coordination Meeting, pp. 171–81. Cotonou, Benin:IITA.

Pavis, C. and Lemaire, L. (1997) Resistance of Musa germplasmto the banana weevil borer, Cosmopolites sordidus Germar(Coleoptera: Curculionidae): A review. Infomusa 6, 3–9.

Pavis, C. and Minost, C. (1993) Banana resistance to the bananaweevil borer Cosmopolites sordidus: Role of pseudostemattractivity and physical properties of the rhizome. In J. Ganry(ed.) Breeding Banana and Plantain for Resistance to Diseasesand Pests, pp. 129–42. Montpellier, France: CIRAD.

PCARRD (1988) The Philippines Recommends for Banana, 136pp. PCARRD Technical Bulletin Series No. 66. Los Banos,Philippines.

Peasley, D.L. and Treverrow, N. (1986) Count, Cut and Dry:A Banana Weevil Borer Management Program, 4 pp. Infor-mation from the Department of Agriculture New South Wales.

Pena, J.E. and Duncan, R. (1991) Preliminary results on biologicalcontrol of Cosmopolites sordidus in Florida. Trop. Fruit NewsAug., 8–10.

Pena, J.E, Duncan, R. and Martin, R. (1993) Biological control ofCosmopolites sordidus in Florida. In C.S. Gold and B. Gemmill(eds) Biological and Integrated Control of Highland Bananaand Plantain Pests and Diseases. Proceedings of a ResearchCoordination Meeting, pp. 124–39. Cotonou, Benin: IITA.

Pena, J.E., Gilbin-Davis, R.M. and Duncan, R. (1995) Impactof indigenous Beauveria bassiana (Balsamo) Vuillemin onbanana weevil and rotten sugarcane weevil (Coleoptera:Curculionidae) populations in banana in Florida. J. Agric.Entomol. 12, 163–7.

Perfecto, I. (1994) The transformation of Cuban agriculture afterthe cold war. Am. J. Alter. Agric. 9, 98–108.

Perfecto, I. and Castineiras, A. (1998) Development of thepredaceous ants and their conservation in agroecosystems. InP. Barbosa (ed.) Conservation Biological Control, pp. 269–89.San Diego: Academic Press.

Persley, G.J. and de Langhe, E.A. (eds) (1987) Bananaand plantain breeding strategies. Proc. inter. workshop,Cairns, Australia 13–17 October 1986, ACIAR and INIBAP.Montpellier, France.

Pianka, E.R. (1970) On r- and K-selection. Am. Nat. 104, 592–7.Pinese, B. (1989) Controlling banana weevil borer. Ban. Bull.

Aust. 53(1), 6,8.Pinese, B. and Piper, R. (1994) Bananas: Insect and Mite Manage-

ment, 67 pp. Department of Primary Industries, Queensland,Australia.

Pinto, A.P.D. (1928) The two weevil pests of plantains (Musasapientum L.): Cosmopolites sordidus Germ. and Odoiporuslongicollis Oliv. Trop. Agric. 70, 216–24.

Pone, S. (1994) Status of nematode and weevil borer problemsin some of the Pacific islands. In: R.V. Valmayor, R.G Davide,J.M. Stanton, N.L. Treverrow and V.N. Roa (eds) Proc. BananaNematode/Borer Weevil Conf. Kuala Lumpur, 18–22 April1994, pp. 90–105. Los Banos, Philippines: INIBAP.

Prando, H.F., Lichtemberg, L.A and Hinz, R.H. (1987) Flutuacaopopulacional da broca da bananeira. Pesq. Andam. 74, 1–3.

Prasad, J.S. and Seshu Reddy, K.V. (1994) Hot water treatmentfor banana planting material made easier. Infomusa 3(2), 16.

Price, N.S. (1993) Preliminary weevil trapping studies inCameroon. In C.S. Gold and B. Gemmill (eds) Biological andIntegrated Control of Highland Banana and Plantain Pests andDiseases. Proceedings of a Research Coordination Meeting,pp. 57–67. Cotonou, Benin: IITA.

Price, N.S. (1994) Alternate cropping in the management ofRadopholus similis and Cosmopolites sordidus, two importantpests of banana and plantain. Inter. J. Pest Manage. 40, 237–44.

Price, N.S. (1995a) The origin and development of banana andplantain cultivation. In: S. Gowen (ed.) Bananas and Plantains,pp. 1–12. London: Chapman and Hall.

Price, N.S. (1995b). The use of a modified pseudo-stem trap-ping technique for assessing the efficacy of insecticides againstbanana-borer weevil. Fruits 50, 23–6.

Prior, C., Jollands, P. and Le Patourel, G. (1988) Infectivity of oiland water formulations of Beauveria bassiana (Deuteromy-cotina:Hyphomycetes) to the cocoa weevil pest Panthorhytesplutus (Coleoptera: Curculionidae). J. Invertebr. Pathol. 52,66–72.

Pulido, J. (1982) Estudios sobre Cosmopolites sordidus Germar:Plaga del platano. Congreso Socolen, Colombia, p. 37.

Pulido, J. (1983) Manejo del picudo negro del platano.Experimento: Ciclo de vida del picudo negro del platano(Cosmopolites sordidus Germar) (Coleoptera: Curculionidae).Proyecto No. 10. ICA-Palmira. Colombia, p. 3.

Pullen, J. (1973) The control of the banana weevil (Cosmopolitessordidus) in Latin America and the Caribbean with Pirimiphos-Ethyl. PANS 19, 178–81.

Rajamony, L., George, K.C., Anitha, N. and Radhakrishnan, T.C.(1993) Stability and adaptation of banana clones belong toAAA group. Planter 69, 343–53.

Rajamony, L., George, K.C., Anitha, N. and Radhakrishnan, T.C.(1994) Assesment of banana (Musa xparadiaca) clones of


AAB group based on stability and adoption. Indian J. Agric.Sci. 64, 521–6

Rajamony, L., Anitha, N., Radkhakrishnan, T.C. and George, K.C.(1995) Variability of yield and yield components in the ABBgroup of bananas. Planter 71, 161–8.

Reinecke, D. (1976) Distribucion del ‘picudo negro’ del platano(Cosmopolites sordidus) en Cuba. Revista Especial Diez Anosde Collaboration Cientifica CUBA-RDA, pp. 52–6. INIFAT.

Reyes-Rivera, H. (2000) Volatile Semiochemicals for Biologi-cal Control of Cosmopolites sordidus. Progress reports of spe-cial grants: Tropical Agriculture. Grant No. 92-34135–U6518.University of Puerto Rico. Rio Piedras.

Risch, S.J., Andow, D. and Altieri, M.A. (1983) Agroecosystemdiversity and pest control: Data, tentative conclusions and newresearch directions. Environ. Entomol. 12, 625–9.

Robalino, G., Roman, J. and Cordero, M. (1983) Efecto delnematicida-insecticida oxamil aplicado al suelo y a las axilasde las hojas del bananero. Nematropica 13, 135–43.

Roberts, F.S. (1955) The banana root borer (Cosmopolitessordidus Germ.), 11 pp. Unpublished manuscript. United FruitCompany. La Lima, Honduras.

Roberts, F.S. (1958) Insects affecting banana production inCentral America. Proc. Tenth Int. Congr. Entomol. 3, 411–15.

Roberts, F.S., Flynn, J.E. and Thornton, N.C. (1955) Control ofthe banana root weevil, Cosmopolites sordidus, in Honduras, 11pp. Unpublished manuscript. United Fruit Company. La Lima,Honduras.

Roche, R. (1975) Comunicacion preliminar sobre la hormigaTetramorium guineense para el control biologico del picudonegro del platano. Rev. Agric. (Cuba) 8, 35–7.

Roche, R. and Abreu, S. (1982) Dispersion de la hormigaTetramorium guineense (Mayr) (Hymenoptera: Formicidae).Cienc. Agric. 13, 122.

Roche, R. and Abreu, S. (1983) Control del picudo negro delplatano (Cosmopolites sordidus) por la hormiga Tetramoriumguineense. Cienc. Agric. 17, 41–9.

Roche, R. and Perez, M.F. (1985) Patron de actividad del formi-cido Tetramorium guineense en Cuba. Cienc. Agric. 24, 30–4.

Rodriguez, J.C. (1989) Seleccion y desinfestation de cormos parala siembra de platano en Tabasco, 4 pp. Instituto Nacional deInvestigaciones Forestales y Agropecuarias.

Roman, J., Oramas, D., Green, J. and Torres, A. (1983) Control ofnematodes and black weevils in plantain. J. Agric. Univ. PuertoRico 67, 270–7.

Root, R.B. (1973) Organization of a plant–arthropod associationin simple and diverse habitats: the fauna of collards (Brassicaoleracea). Ecol. Monogr. 43, 95–124.

Roth, L. and Willis, E. (1963) The humidity behavior ofCosmopolites sordidus Germar (Coleoptera:Curculionidae).Ann. Entomol. Soc. Am. 56, 41–2.

Roy, R.S. and Sharma, C. (1952) Diseases and pests of bananasand their control. Indian J. Hortic. 9, 39–52.

Rukazambuga, N.D.T.M. (1996) The Effects of Banana Weevil(Cosmopolites sordidus Germar) on the Growth and Produc-tivity of Bananas (Musa AAA EA) and the Influence of HostVigour on Attack, 249 pp. Ph.D. thesis, University of Reading.United Kingdom.

Rukazambuga, N.D.T.M., Gold, C.S. and Gowen, S.R. (1998)Yield loss in East African highland banana (Musa spp.,AAA-EA group) caused by the banana weevil, Cosmopolitessordidus Germar. Crop Prot. 17, 581–9.

Rukazambuga, N.D.T.M., Gold, C.S. and Gowen, S.R. (2002).The influence of crop management on banana weevil,Cosmopolites sordidus (Coleoptera: Curculionidae) popula-tions and yield of highland cooking banana (cv Atwalira) inUganda. Bull. Entomol. Res.: 92, 413–21.

Rwekika, E. (1996) Feeding Allelochemicals for the BananaWeevil Cosmopolites sordidus Germar, 126 pp. Ph.D. thesis,University of Dar-es-Salaam, Tanzania.

Rwekika, E., Ndiege, I.O., Hassanali, A., Lwande, W. andMhehe, G. (2003) Identification of some of the major feed-ing stimulants for the banana weevil Cosmopolites sordidus.J. Chem. Ecol.: In press.

Salazar, A. A. (1999) Efecto de Mucuna deeringiana (BORT)Merr. sobre el picudo del cormo, Cosmopolites sordidusGermar (Coleoptera: Curculionidae) en platano, 32 pp. Mas-ters thesis. University of Puerto Rico.

Sampaio, A.S., Myazaki, I., Suplicy Filho, N. and Oliveira, D.A.(1982) ‘Broca da banaeira’ – Cosmopolites sordidus (Germar,1824) (Coleoptera: Curculionidae) resistente ao aldrin esue controle com inseticidas sistemicos aplicados no solo.Biologico 48, 91–8.

Sarah, J.L. (1990) Les charancons des bananiers. Fruits (Specialissue – Bananas) 68–71

Sarah, J.L. (1994) CIRAD-FLHOR research actions on nema-todes and black borer weevil of bananas and plantains, 11 pp.Unpublished MSS. CIRAD-FLHOR. Montpellier, France.

Saraiva, A. (1964) O gorgulho da bananeria Cosmopolitessordidus (Germar) no arquipelago de Cabo Verde. Garc. deOrta 12, 241–9.

Schill, P. (1996) Final Report: Distribution, Economic Sta-tus, Ecology and Biological Control of Plantain Pests andDiseases in West and Central Africa with Emphasis on theWeevil Cosmopolites sordidus (Germar). July 1993–June 1996.Cotonou, Benin: IITA.

Schill, P., Afreh-Nuamah, K., Gold, C., Ulzen-Apiah, F.,Paa Kwesi, E., Peprah, S.A. and Twumasi, J.K. (1997) FarmersPerception of Contraints in Plantain Production in Ghana, 41pp. + maps. Plant Health Management Division MonographNo. 5. Ibadan, Nigeria: IITA.

Schmidt, C.T. (1965) O gorgulho da bananeira em Sao Tome.Estudos Agronomicos 6, 97–104.

Schmidt, F.H. and Lauer, W.L. (1977) Developmental polymor-phism in Choristoneura spp. (Lepidoptera: Torticidae). Ann.Entomol. Soc. Am. 70, 750–6.

Schmitt, A.T. (1993) Biological Control of the Banana Weevil(Cosmopolites sordidus (Germar)) with Entomogenous Nema-todes, 210 pp. Ph.D. thesis. University of Reading, UK.

Schmitt, A.T., Gowen, S.R. and Hague, N.G.M. (1992) Baitingtechniques for the contol of Cosmopolites sordidus Germar(Coleoptera: Curculionidae) by Steinernema carpocapsae(Nematoda): Steinernematidae). Nematropica 22, 159–63.

Schoeman, P.S. and Schoeman, M.H. (1999) Transmissionof Beauveria bassiana from infected to uninfected adultsof the banana weevil Cosmopolites sordidus (Coleoptera:Curculionidae). Afr. Plant Prot. 5, 53–4.

Sebasigari, K. and Stover, R.H. (1988) Banana Diseases and Pestsin East Africa: Report of a Survey in November 1987, 15 pp. +appendices and tables. Montepellier, France: INIBAP.

Sein, F. Jr. (1934) Paring and heat sterilization of the cormsto eliminate the banana root weevil Cosmopolites sordidusGermar. J. Agric. Univ. Puerto Rico 18, 411–16.


Sen, A.C. and Prasad, D. (1953) Pests of banana in Bihar. IndianJ. Entomol. 15, 240–6.

Sengooba T. (1986) Survey of Banana Pest Problem Complexin Rakai and Masaka Districts, August 1986: PreliminaryTrip Report, 10 pp. Namulonge Research Station, Namulonge,Uganda. Unpubl.

Sery, G.D. (1988) Oreintationes de recherches pour la mise aupoint de nouvelles methodes de lutte contre les nematodes et alcharancon du bananier et du bananier plantain. In Nematodesand the Borer Weevil in Bananas: Proceedings of a Work-shop Bujumbura, Burundi, 7–11 December 1987, pp. 83–5.Montpellier, France: INIBAP.

Seshu Reddy, K.V. and Lubega, M.C. (1993) Evaluation of bananacultivars for resistance/tolerance of the weevil Cosmopolitessordidus Germar. In J. Ganry (ed.) Breeding Banana andPlantain for Resistance to Diseases and Pests, pp. 143–148.Montpellier, France: CIRAD.

Seshu Reddy, K.V., Koppenhofer, A.M. and Uronu, B. (1993) Cul-tural practices for the control of the banana weevil. In C.S. Goldand B. Gemmill (eds) Biological and Integrated Control ofHighland Banana and Plantain Pests and Diseases. Proceed-ings of a Research Coordination Meeting, pp. 140–7. Cotonou,Benin: IITA.

Seshu Reddy, K.V., Prasad, J.S., Ngode, L. and Sikora, R.A.(1995) Influence of trapping of the banana weevil,Cosmopolites sordidus (Germar 1824) on root-lesion nema-tode, Pratylenchus goodeyi (Sher and Allen 1953) populationdensities and subsequent banana yield. Acta Oecol. 16, 593–8.

Seshu Reddy, K.V., Prasad, J.S. and Sikora, R.A. (1998) Bioin-tensive management of crop borers of banana. In S.K. Saini(ed) Proceed. Symp. Biol. Control Trop. Crop Habitats: ThirdInt. Conf. Trop. Entomol., 30 October–4 November 1994,pp. 261–87. Nairobi, Kenya: ICIPE Science Press.

Shanahan, G.J. and Goodyer, G.J. (1974) Dieldrin resistance inCosmopolites sordidus in New South Wales, Australia. J. Econ.Entomol. 67, 446–7.

Shell (1967) Combate de Plagas. In Cambures. Serie A. No. 29.pp. 20–37. Cagua, Venezuela.

Shillingford, C.A. (1988) Review of parameters used for eval-uating nematode and borer damage in bananas and plantains.In Nematodes and the Borer Weevil in Bananas: Proceedingsof a Workshop, Bujumbura, Burundi, 7–11 December 1987,pp. 87–90. Montpellier, France: INIBAP.

Sikora, R.A., Bafokuzara, N.D., Mbwana, A.S.S., Oloo, G.W.,Uronu, B. and Seshu Reddy, K.V. (1989) Interrelationshipbetween banana weevil, root lesion nematode and agronomicpractices, and their importance for banana decline in the UnitedRepublic of Tanzania. FAO Plant Prot. Bull. 37, 151–7.

Silva, S. de O. and Fancelli, M. (1998) Banana insect pests. InV. Galan (ed) Proc. Int. Symp. Banana Subtrop., pp. 385–93.Tenerife, Spain.

Simmonds, N.W. (1966) Bananas, 512 pp. London: LongmansPress.

Simmonds, N.W. and Shepherd, K. (1955) The taxonomy andorigins of the cultivated bananas. J. Linn. Soc. 55, 302–12.

Simmonds, N.W. and Simmonds, F.J. (1953) Experiments on thebanana borer, Cosmopolites sordidus in Trinidad, B.W.I. Trop.Agric. 30, 216–23.

Simon, S. (1993) Pests of bananas in the French West Indies.Infomusa 2(1), 8.

Simon, S. (1994) La lutte integree contre le charancon noir desbananiers Cosmopolites sordidus. Fruits 49, 151–62.

Singh, J.P. (1970) Insect pests of banana. Allah. Farmer 44,295–303.

Smith, D. (1995) Banana weevil borer control in south-easternQueensland. Aust. J. Exp. Agric. 35, 1165–72.

Soares, G., Figueiredo, J., Lopes, H., Lopes, J. and Mello, R.J.(1980) Controle de pragas da baneneira (Musa sp.) com o fungoentomogeno Beauveria bassiana (Bals.) Vuill. Pesq. agropec.pernamb. Recife 4, 149–55.

Sotomayor, B. (1972) Resistencia de Cosmopolites sordidusGermar a los compuestos organoclorados en el Ecuador. Rev.Peru. Entomol. 15, 169–75.

Speijer, P.R., Budenberg, B. and Sikora, R.A. (1993) Rela-tionships between nematodes, weevils, banana and plantaincultivars and damage. Ann. Appl. Biol. 123, 517–25.

Speijer, P.R., Gold, C.S., Kajumba, C. Karamura, E.B. (1995)Nematode infestation of ‘clean’ banana planting material infarmers fields in Uganda. Nematologica 41, 344.

Sponagel, K.W., Diaz, F.J. and Cribas, A. (1995) El picudo negrodel platano, Cosmopolites sordidus Germar, 35 pp. + plates.La Lima, Honduras: FHIA.

Ssennyonga, J.W., Bagamba, F., Gold, C.S., Tushemereirwe, W.K.,Ssendege, R. and Katungi, E. (1999) Understanding CurrentBanana Production with Special Reference to Integrated PestManagement in Southwestern Uganda, 47 pp. Nairobi, Kenya:ICIPE.

Stanton, J.M. (1994) Status of nematode and weevil borerproblems in Australia. In R.V. Valmayor, R.G Davide,J.M. Stanton, N.L. Treverrow, and V.N. Roa (eds) Proc. BananaNematode/Borer Weevil Conf., Kuala Lumpur, 18–22 April1994, pp. 48–56. Los Banos, Philippines: INIBAP.

Staver, C. (1989) Why farmers rotate fields in maize-cassava-plantain bush fallow agriculture in the West Peruvian Amazon.Hum. Ecol. 17, 401–26.

Stover, R.H. and Simmonds, N.W. (1987) Bananas: ThirdEdition, 469 pp. New York: John Wiley and Sons.

Stephens, C.S. (1984) Notes of three Philicoptus banana pests(Coleoptera: Curculionidae) and notes on other weevils inMindanao, Philippines. Philip. Agric. 67, 243–53.

Sumani, A.J. (1997) Patterns of Relationship Between Banana(Musa spp.) Types and the Banana Weevil, Cosmopolitessordidus (Germar) (Coleoptera: Curculionidae), 118 pp.Ph.D. thesis, University of Zambia. Lusaka, Zambia.

Suplicy Filho, N. and A.S. Sampaio. (1982) Pragas da bananeira.Biologico 47, 169–82.

Swaine, G. and Corcoran, R. (1973) A field trial on a suspecteddieldrin-resistant population of banana weevil borer. Queensl.J. Agric. Anim. Sci. 30, 79–83.

Swaine, G., Pinese, B. and Corcoran, R. (1980) Dieldrin resis-tance in the banana weevil borer, Cosmopolites sordidus, Germ.in Queensland. Queensl. J. Agric. Anim. Sci. 37, 35–7.

Swaine, R.B. (1952) Insect problems in Nicaragua. FAO PlantProt. Bull. 1, 27–8.

Swennen, R., Wilson, G.F. and D. Decoene, D. (1988) Priori-ties for future research on the root system and corm in plan-tains and bananas in relation with nematodes and the bananaweevil. In Nematodes and the Borer Weevil in Bananas: Pro-ceedings of a Workshop Bujumbura, Burundi, 7–11 December1987, pp. 91–6. Montpellier, France: INIBAP.

Taylor, B. (1991) Research field work on upland bananas,Musa spp., principally acuminata triploid AAA types in theKagera region of Tanzania. With observations on growth and


causes of decline in crop yield. Riv. Agric. Subtropi. Tropic. 85,349–92.

Tezenas du Montcel, H. (1987) Plantain Bananas. The TropicalAgricultural Series, 106 pp. London: MacMillan Co.

Tinzaara, W., Karamura, E. and Tushemereirwe, W. (1999a)Preliminary observations on natural enemies associated withthe banana weevil Cosmopolites sordidus Germar in Uganda.Infomusa 8(1), 28–9.

Tinzaara, W., Tushemereirwe, W. and Kashaija, I. (1999b) Thepotential for using pheromone traps for the control of bananaweevil Cosmopolites sordidus Germar in Uganda. In E. Frison,C.S. Gold, E.B. Karamura and R.A. Sikora (eds). MobilizingIPM for Sustainable Banana Production in Africa. Proceedingsof a Workshop on Banana IPM, Nelspruit, South Africa, 23–28November 1998, pp. 327–32. Montpellier, France: INIBAP.

Traore, L. (1995) Facteurs biologiques de mortalite decurculionidae en mileux tempere et tropical, 192 pp. Ph.D.thesis, McGill University, Montreal, Canada.

Traore, L., Gold, C.S., Boivin, G. and Pilon, J.G. (1996) Devel-oppement postembryonnaire du charancon du bananier. Fruits51, 105–13.

Traore, L., Gold, C.S., Pilon, J.G. and Boivin, G. (1993) Effectsof temperature on embryonic development of banana weevil,Cosmopolites sordidus Germar. Afr. Crop Sci. J. 1, 111–6.

Trejo, J.A. (1969). El picudo negro del banano, Cosmopolitessordidus (Germar). Agricultura en El Salvador 9, 22–5.

Treverrow, N. (1985) Banana Weevil Borer. Agfacts, 3 pp. Depart-ment of Agriculture, New South Wales, Australia.

Treverrow, N. (1993) An Integrated Management Program forBanana Weevil Borer. Final Report. HRDC. Project No.Fr/0012/RO, 40 pp. Wollongbar New South Wales, Australia:Wollongbar Agric. Institute.

Treverrow, N.L. (1994) Control of the banana weevil borer,Cosmopolites sordidus (Germar) with entomopathogenicnematodes. In R.V. Valmayor, R.G Davide, J.M. Stanton,N.L. Treverrow and V.N. Roa (eds) Proc. Banana Nematode/Borer Weevil Conf., Kuala Lumpur, 18–22 April 1994,pp. 124–38. Los Banos, Philippines: INIBAP.

Treverrow, N.L. and Bedding, R.A. (1993) Development of a sys-tem for the control of the banana weevil borer, Cosmopolitessordidus with entompathogenic nematodes. In R. Bedding,R. Akhurst and H. Kaya. (eds) Nematodes and the Biologi-cal Control of Insect Pests, pp. 41–7. Melbourne, Australia:CSIRO.

Treverrow, N. and Maddox, C. (1993) The distribution ofCosmopolites sordidus (Germar) (Coleoptera: Curculionidae)between various types of banana plant material in relation tocrop hygiene. Gen. Appl. Entomol. 23, 15–20.

Treverrow, N., Bedding, R., Dettmann, E.B. and Maddox,C. (1991) Evaluation of entomopathogenic nematodes forthe control of Cosmopolites sordidus Germar (Coleoptera:Curculionidae), a pest of bananas in Australia. Ann. Appl. Biol.119, 139–45.

Treverrow, N., Peasley, D. and Ireland, G. (1992) Banana WeevilBorer: A Pest Management Handbook for Banana Growers, 28pp. Banana Industry Committee, New South Wales Agriculture.

Tsai, Y.P. (1986) Major insect pests of banana. In Plant Protectionin the Republic of China (1954–1984), 6 pp. Taiwain BananaResearch Institute.

Turner, D.W. (1994) Bananas and plantains. In B. Schaffer andP.C. Anderson (eds) Handbook of Environmental Physiology

of Fruit Crops. Volume III, Subtropical and Tropical Crops,pp. 37–61. Boca Raton, Florida: CRC Press.

Tushemereirwe, W.K., Kashaija, I.N., Tinzaara, W., Nankinga, C.and New, S. (2000)Banana Production Manual: A Guide toSuccessful Banana Production in Uganda, 77 pp. Kampala,Uganda: NARO and ADC IDEA Project.

Udzu, A. (1997) Study of the Effects of Banana Weevil and Nema-todes on the Growth and Yield of Plantain (Musa AAB), 83 pp.Masters thesis, University of Ghana, Legon.

Uronu, B.E.M.A. (1992) The Effect of Plant Resistance and Cul-tural Practices on the Population Densities of Banana WeevilCosmopolites sordidus (Germar) and on Banana Yield, 216 pp.Ph.D. thesis, Kenyatta University, Nairobi, Kenya.

Uzakah, R.P. (1995) The Reproductive Biology, Behaviour andPheromones of the Banana Weevil, Cosmopolites sordidusGermar (Coleoptera: Curculionidae), 177 pp. Ph.D. thesis,University of Ibadan. Nigeria.

Valentine, B.D. and Valentine, B.S. (1957) Some injurious insectsin Haiti. Coleopt. Bull. 11, 29–32.

Valmayor, R.V., Davide, R.G., Stanton, J.M., Treverrow, N.L.and Roa, V.N. (eds). (1994) Proc. Banana Nematode/BorerWeevil Conf., Kuala Lumpur, 18–22 April 1994. Los Banos,Philippines: INIBAP.

Van den Enden, H. and Garcia, E.A. (1984) Reconocimiento delcontrol natural del picudo negro del platano (Cosmopolitessordidus Germar) en la zona del gran Caldas, 114 pp. Univer-sidad Nacional, Facultad de Agronomia. Manizales, Colombia.

van Driesche, R.G. and T.S. Bellows Jr. (1996) BiologicalControl, 539 pp. New York: Chapman and Hall.

Varela, A.M. (1993) Report on a trip to Bukoba and Mulebadistricts (Kagera region) to assess the possibility for biolog-ical control of banana weevil with ants, 15 pp. Unpublishedmanuscript. Agricultural Research Institute Maruku, Tanzania.

Veitch, R. (1929) The banana weevil borer. In R.Veitch andJ.H. Simmonds (eds) Pests and Diseases of Queensland Fruitsand Vegetables, pp. 255–63.

Velasco, P. (1975) Incidencia y control quimico del picudo negrodel platano Cosmopolites sordidus (Germar). Agric. Tecn. Mex.3, 361–4.

Viana, A.M.M. (1992) Comportamento de Agregaca e acasaole-mento de Cosmopolites sordidus (Coleoptera: Curculionidae),mediado por semoquimcios, em olfactometro, 75 pp.Masters thesis, Vicosa, Brazil.

Viana, A.M.M. and Vilela, E.F. (1996) Comportamiento de cortee acasalamento de Cosmopolites sordidus Germar (Coleoptera:Curculionidae). An. Soc. Entomol. Brasil 25, 347–50.

Vilardebo, A. (1950) Conditions d’un bon rendement du peigeagede Cosmopolites sordidus. Fruits 5, 399–404.

Vilardebo, A. (1960) Los insectos y nematodos de las bananerasdel Ecuador, 78 pp. Instituto Franco-Euaotiano de Investiga-ciones Agromicas, Paris.

Vilardebo, A. (1967) Resistance of the banana tree weevilCosmopolites sordidus (Germar) to chlorinated hydrocarboninsecticides. Int. Congr. Plant Prot. 586.

Vilardebo, A. (1973) Le coefficient d’infestation, critered’evaluation du degre d’attaques des bananeraies parCosmopolites sordidus Germ. le charancon noir du bananier.Fruits 28, 417–31.

Vilardebo, A. (1977) Crop Loss Assessments, 3 pp. FAO/CABI.Supplement 2. Rome.


Vilardebo, A. (1984) Problemes scientifiques poses parRadopholus similis et Cosmopolites sordidus en culturesbananieres des zones francophones de production. Fruits 39,227–33.

Viswanath, B.N. (1976) Studies on the Biology, Varietal Responseand Control of Banana Rhizome Weevil, Cosmopolites sordidus(Germar) (Coleoptera: Curculionidae), 152 pp. Ph.D. thesis,University of Bangalore. Bangalore, India.

Viswanath, B. (1981) Development of Cosmopolites sordidus(Coleoptera: Curculionidae) on banana varieties in South India.Colemania 1, 57–8.

Vittayaruk, W., Wattanachaiyingcharoen, W., Chuaycharoen, T.and Wattanachaiyingcharoen, D. (1994) Status of weevil borerproblems affecting banana in Thailand. In R.V. Valmayor,R.G. Davide, J.M. Stanton, N L. Treverrow and V.N. Roa(eds) Proc. Banana Nematode/Borer Weevil Conf., KualaLumpur, 18–22 April 1994, pp. 106–114. INIBAP. Los Banos,Philippines.

Walangululu, M., Litucha, B.M. and Musasa, M. (1993) Potentialfor the control of the banana weevil Cosmopolites sordidusGermar with plants reputed to have an insecticidal effect.Infomusa 2, 9.

Walker, A.K. and Dietz, L.L. (1979) A review of entomophagousinsects in the Cook Islands. N Z Entomol. 7, 70–82.

Walker, P.T., Hebblethwaite, M.J., and Bridge, J. (1983) Projectfor Banana Pest Control and Improvement in Tanzania:A Report to the Government of Tanzania, 141 pp. + maps.Tropical Development Research Institute, London.

Wallace, C.R. (1938) Measurement of beetle borer migration inbanana plantations. J. Aust. Inst. Agric. Sci. 4, 215–19.

Wardlaw, C.W. (1972) Appendix III. In Banana Diseases Includ-ing Plantains and Abaca, 2nd Edition, pp. 567–70. London:Longman Group Ltd.

Wardrop, D.J. and Forslund, R.E. (2002) Synthesis of(+−)-episordidin. Tetrahedron Lett. 43, 737–9.

Waterhouse, D.F. (1993) The Major Arthropod Pests and Weedsof Agriculture in Southeast Asia. Canberra, Australia: ACIAR,141 pp.

Waterhouse, D.F. and Norris, K.R. (1987) Cosmopolites sordidus(Germar). In D.F. Waterhouse and K.R. Norris (eds) BiologicalControl: Pacific Prospects, pp. 152–8. Melbourne, Australia:Inkata Press.

Waterhouse, D.F. and Sands, D.P.A. (2001) Classical Biologi-cal Control of Arthropods in Australia. Canberra, Australia:CSIRO.

Weddell, J.A. (1932) The banana weevil borer: Brief notes onPlaesius javanus Er., the histerid predator. Queensl. Agric. J.38, 19–25.

Weddell, J.A. (1945) The banana weevil borer. Queensl. Agric. J.51, 85–91.

Whalley, P. (1957) The banana weevil and its control. East Afr.Agric. J. 23, 110–12.

Williams, D.B., Laville, B. and Fagan, H.J. (1986) ImprovingWindward Islands banana production through phytosanitation.In Sem. Proc. Improv. Citrus Banana Prod. Carib. Phytosanit.,2–5 December 1986, Paises Bajos, Ede. CTA. pp. 14–20.Windward Island Banana Association, St Lucia.

Wolcott, G.N. (1924) The food of Porto Rican Lizards. J. Dep.Agric. Puerto Rico VII(4), 5–37.

Wolcott, G.N. (1948) The insects of Puerto Rico. J. Agric. Univ.Puerto Rico 32, 101–45

Wolfenbarger, D. (1964) Banana root borer and its control inFlorida. Proceedings Caribbean. Region, Am. Soc. Hortic. Sci.8, 67–70.

Woodruff, R.E. (1969) The Banana Root Borer (Cosmopolitessordidus (Germar)) in Florida (Coleoptera: Curculionidae).Florida Department of Agricultural Consumer Services.Entomological Circular No. 88. 2 pp.

Wright, W. (1977) Insecticides for the control of dieldrin-resistantbanana weevil borer, Cosmopolites sordidus Germar. Aust. J.Exp. Agric. Anim. Husband. 17, 499–504.

Yaringano, C. and van der Meer, F. (1975) Control del gorgojodel platano, Cosmopolites sordidus Germar, mediante tram-pas diversas y pesticidas granulados. Rev. Peru. Entomol. 18,112–16

Ysenbrandt, H., Fogain, R. and Messiaen, S. (2000) Infestationlevels of weevil species on Musa cultivars Grande Naine (AAA)and French Sombre (AAB) and subsequent plant mortality inCameroon. Afr. Plant Prot. 6, 21–4.

Zar, J.H. (1984) Biostatistical Analysis. 718 pp. Englewood Cliffs,New Jersey: Prentice-Hall, Inc.

Zem, A.C. and Alves, E.J. (1976) A broca da bananieraCosmopolites sordidus (Germar, 1824) no Estado da Bahia –1 – Incidencia e movimentacao: In Congr. Brasi. Frutic., 5,pp. 284–9. Pelotas-R.S.

Zem, A.C., Rodrigues, J.A.S. and Alves, E.J. (1978) Com-portamento do cultivares de bananeira (Musa spp). aoataque do Broca do Rizoma (Cosmopolites sordidus Germar)(Coleoptera: Curculionidae). Ecosistema 3(3), 8–10.

Zimmerman, E.C. (1968a) The Cosmopolites banana weevils(Coleoptera: Curculionidae; Rhynchophorinae). Pacific Insects10, 295–9.

Zimmerman, E.C. (1968b) Rhynchophorinae of southeasternPolynesia. Pacific Insects 10, 47–77.

Zimmerman, E.C. (1968c) Cosmopolites pruinosus, a new pestof banana. J. Econ. Entomol. 61, 870–1.

Biology and integrated pest management for the banana weevil

Documents

weevil damage

banana weevils

weevil attack

weevil pest status

weevilfree material

banana breeding

type of banana

banana mats