Chapter 6 Chemical control of Cosmopolites sordidus in South Africa

Transcription

1 Chpter 6 Chemicl control of Cosmopolites sordidus in South Afric 188

2 Abstrct Cosmopolites sordidus (bnn weevil) is mjor production constrint in most res where bnns re grown. The weevil is difficult to control, nd chemicl control rgubly provides the best opportunity to mnge the pest. The im of this study ws to determine the efficcy of injecting bifenthrin, chlorpyrifos, fipronil, imidcloprid, oxmyl nd wter (control) into residul bnn plnts. The chemicls were dministered every even numbered month over 2 yers t two loctions in Southern KwZulu-Ntl, South Afric. Yield, weevil dmge nd pseudostem girth of plnts felled from August to October were mesured, while dult beetle densities were ssessed over 4 weeks in October nd April. Nemtode smples were nlysed in October every yer. Dmge prmeters included the Coefficient of Infesttion, the Percentge Coefficient of Infesttion (PCI) t two intervls, the summed PCI vlue, the percentge cross sectionl dmge of the centrl cylinder nd cortex, nd the men cross sectionl dmge percentge. Replicted block designs were used in the experiments. The prmeters were similr before the onset of the tril. Fruit yield nd plnt girth, corrected by nemtode densities, were not significntly incresed fter chemicl pplictions, nor were the nemtodes controlled. Fipronil nd imidcloprid were highly effective ginst C. sordidus, minimising dmge to the periphery, cortex nd centrl cylinder of the rhizome nd significntly reduced dult density. Fipronil cused 95% nd imidcloprid 100% reduction in the cross sectionl dmge of the centrl cylinder, the dmge prmeter most closely relted to yield. Injection of fipronil nd imidcloprid provides n optiml chemicl strtegy in n integrted pest mngement progrmme for the bnn weevil. Keywords: Insecticide, injection, yield, dmge, bnn weevil 189

3 6.1 Introduction The bnn weevil, Cosmopolites sordidus (Germr), is n importnt pest of Mus nd Ensete (Stover & Simmonds 1987, Gold & Messien 2000; Gold et l. 2003) nd the dominnt insect pest of bnns in South Afric. Adults feed on plnt tissues or crop debris but the dmge inflicted is negligible (Frnzmnn 1972; Treverrow et l. 1992). Femles oviposit their eggs singulrly (Froggtt 1925, Simmonds 1966; Frnzmnn 1972) in the crown of the rhizome nd pseudostem bse (Aber et l. 1999), fvouring flowered plnts (Treverrow et l. 1992; Aber et l. 1999). Upon emergence, the lrve tunnel into the rhizome, producing distinctive circulr, debrisfilled tunnels (Frnzmnn 1972). Interior corm dmge ffects nutrient trnsport nd stem growth (Tylor 1991), while peripherl dmge dversely ffects root development (Gold et l. 1994). The pup develops in chmber t the corm periphery (Frnzmnn 1972) nd eclosion produces reddish, brown dult (tenerl stge), which lter becomes uniformly dull blck (Pinese & Elder 2004). Infested plnts show stunted growth, delyed mturtion (Gold et l. 1998), reduced bunch weight, nd cn snp or topple (Btchelder 1954, Frnzmnn 1972; Koppenhöfer 1993). Infesttion by bnn root nemtodes cn show similr symptoms, including reduction in vigour, lef chlorosis, plnt toppling nd yield reduction (Bujulu et l. 1983, Smith 1995; Willers et l. 2001). Chemicl control of the beetle hs been employed since the erly 20th century. Pesticides consisted minly of Pris Green, followed by the use of orgnochlorines like BHC nd DDT (Froggtt 1925, Cuillé 1950, Simmonds 1966; Treverrow et l. 1992). The chemicls were usully pplied with flour or other substnces s bits (Froggtt 1925, Cuillé 1950, Simmonds 1966; Treverrow et l. 1992). The method ws not very effective (Simmonds 1966) nd the persistent cyclodienes, dieldrin nd ldrin, showed high efficcy s soil tretment ginst the bnn weevil (Brithwite 1958). Cyclodienes ws used extensively round the world from the mid 1950 s (Edge 1974) nd ws found to be effective for up to 2 yers fter ppliction (Brithwite 1967). Before 1970, however, resistnce to cyclodienes ws widely dignosed (Vilrdebó 1967, Anonymous 1969; Shnhn & Goodyer 1974). Investigtions into lterntive chemicls (minly orgnophosphtes nd crbmtes) showed chlordecone (orgnochlorine), pirimiphos-ethyl, chlorpyrifos, 190

4 prothiophos nd ethoprophos s vible for binnul pplictions, but dizinon ws unsuitble becuse of its short residul ction (Wright 1977, Collins et l. 1991; Smith 1995). Aldicrb, terbufos, crbofurn, crbosulfn, oxmyl, fenmiphos (Román et l. 1979, Cárdens 1984, De Jger et l. 1991, Vittyruk et l. 1994, Chvrri-Crvjl & Irizrry 1997; Fogin et l. 2002), isofenphos, iszofos (Bujulu et l. 1983), phoxim (Nuno & Ribeiro 2002) tebupirimiphos, cdusfos (Quilici 1993), fosthizte (Chbrier et l. 2002), phorte, disulfoton, quinlphos (Viswnth 1977), cephte, diethyl, pd, monocrotophos, deltmethrin (pyrethroid) (Molin 1994), fipronil (phenyl pyrzole) (Price 1995; Fogin et l. 2002) nd bifenthrin (pyrethroid) (Smith 1995) were lso found to be effective. Less thn 10 yers fter widespred orgnophosphte use in Austrli, resistnce to pirimiphos-ethyl, prothiophos, chlorpyrifos nd ethoprophos were reported in Queenslnd nd New South Wles with evidence of cross resistnce to oxmyl but not to crbofurn, iszofos or isofenphos (Collins et l. 1991). Subsequently soil pplictions of bifenthrin were found to be effective, but fipronil, crbosulfn nd furthiocrb were similr to untreted controls in Southest Queenslnd (Smith 1995). Resistnce to crbofurn hs not been found in Ugnd or Austrli (Collins et l. 1991; Gold et l. 1999). The high rte of resistnce development ws ttributed to widespred, regulr pplictions with no popultion monitoring (Collins et l. 1991). Chemicl control with non-systemic pesticides is minly directed ginst dults (Simmonds 1966, Wright 1977; Collins et l. 1991). Dipping corms in insecticide solution were significntly more effective thn hot wter tretment in controlling the weevil in plnting mteril (Crdens Murillo et l. 1986). Chemicl ppliction is commonly recommended in plnting holes (Frnzmnn 1972, Anith et l. 1992; Fogin et l. 2002), to plnt trps (bit sprying) (Treverrow et l. 1992) nd to the bses of bnn plnts (butt sprys) (Brithwite 1958, Bujulu et l. 1983, Collins et l. 1991, Smith 1995; Fogin et l. 2002). In Austrli butt sprys re pplied in spring nd utumn nd chemicls re injected into residul pseudostems during winter (Froggtt 1926, Treverrow 1985, Treverrow et l. 1992, Stnton 1994). Butt sprys re, however, detrimentl to beneficil insects nd only trget dults in close vicinity of plnts (Collins et l. 1991). Bit sprys re pplied to fresh residues every 2nd or 4th week in spring nd utumn. Poison trps sve on insecticide, but re regrded s being reltively ineffective (Simmonds 1959), especilly t high infesttion levels (Treverrow et l. 1992). 191

5 Systemic chemicls (dimethote, omethote, ldicrb, crbofurn, crbosulfn, fenmiphos, fosthizte, iszofos, monocrotophos, oxmyl, phorte nd terbufos) cn potentilly control lrve following uptke by bnn roots fter soil ppliction (Gold et l. 2003). These chemicls provide protective tretment for plnts, but hve reltively shorter residul ctions (Treverrow et l. 1992) nd do not prevent ttcks on plnt residues fter hrvest (Treverrow pers. comm.). Dul ction insecticide-nemticides with systemic ction will be of vlue to tret moderte weevil infesttions when nemtode densities lso require tretment (Treverrow et l. 1992). In South Afric, lte summer nd erly spring butt ppliction of pirimiphosethyl nd ldicrb hs been recommended (Jones & Dieckmnn 1982). Pirimiphosethyl ws used until the mid 1990 s (Schoemn 1996) nd imidcloprid nd prothiofos were used in 1999 (Schoemn et l. 1999). Loclly the pesticides ldicrb, terbufos nd oxmyl were lso reported to be effective in controlling the bnn weevil nd the prtylenchid nemtode, Rdopholus similis (Cobb) (De Jger et l. 1991). Schoemn (1998) reported tht fenmiphos nd cdusfos showed promise to control the weevil in field tril, yet Dochez (1998) showed tht terbufos, fosthizte, ldicrb nd cdusfos did not reduce weevil dmge loclly. Only ldicrb is registered for control of the bnn weevil nd nemtodes in South Afric (Nel et l. 2002; Anonymous 2005). The soil round the plnts is treted nd ppliction is recommended t plnting, during November (lte spring) nd Mrch (lte summer/erly utumn). According to Quilici (1993) nd Schoemn (1998), ldicrb does not provide sufficient control of the weevil nd growers hve lso reported tretment filures. Some desperte growers hve even resorted to illegl nd unregistered chemicl usge (Dochez 1998). The im of this study ws to determine the efficcy of injecting contct nd systemic chemicls into residul bnn mteril in South Afric throughout the yer. 6.2 Mteril nd methods Reserch sites Trils to evlute the efficcy of chemicls ginst the bnn weevil were conducted on two commercil frms t the South Cost of KwZulu-Ntl, South Afric. The tril sites were in Munster (30º59 29 S; 30º14 49 E) nd Rmsgte (30º52 31 S; 30º19 29 E), 72 nd 130 meters bove se level, respectively. Soil in the re is 192

6 Glenros form, with n orthic A nd lithocutnic B zone. It is sndy lom soil with 16% cly, 30% lom nd 54% snd (Dochez 1998). The experiments were conducted from August 2003 to October The loctions were in summer rinfll re ( mm per yer), nd during the trils the mbient temperture rnged from 12 to 25 C. The Cvendish cultivrs Willims nd Chinese Cvendish (AAA group) were grown t the Munster nd Rmsgte trils, respectively. The former ws plnted in November 1995 nd the ltter in November 2000, both t density of 2222 plnts.h -1 ( cm). High mt ws evident in the plnttions, with the collr (junction between pseudostem nd rhizome) commonly more thn 10 cm bove ground level. The Munster plnttion ws drip nd the Rmsgte site sprinkler irrigted with 2 cm wter/week, prctise only suspended if rinfll exceeded tht vlue in the prticulr week. The sites were treted t plnting with the oxime crbmte, ldicrb (Temik 15% GR), t the registered dosge of g..i./mt, to provide nemtode nd weevil control. Regulr chemicl weed control with glyphoste (Roundup), lef removl, desuckering nd propping of bunch bering plnts were prctised. Pre-tril plnt inspections t ll sites reveled rhizome tunnel dmge by C. sordidus. No plnttion hygiene ws prctised nd t both sites ccumulted residues were destroyed in Jnury The sites were reltively similr, but compred to the Rmsgte loction, the older Munster plnttion hd lower plnt density (less cnopy cover) s result of plnt toppling, higher rte of residue desicction, more remnnts present in the field nd not ll residues (pproximtely 65 to 70%) were ttched to the mother plnt Experimentl design Five different chemicls were evluted, but imidcloprid ws only included in the Rmsgte tril (Tble 6.1). Control plnts were injected with wter. Tretments were pplied by injecting 10 ml of chemicl solution (or wter) into residul bnn pseudostems using clibrted knpsck (Clibr stem pplictor, Interlock CC, Pretori, South Afric). The lnce of the bckpck ws end-cpped with spershped dgger, with three injector slits on opposite sides t the distl end (Interlock CC, Pretori, South Afric). Moist tissue of softened, decyed pseudostems (or rhizomes), with t lest distl portion esily compressible by hnd, were injected t 100 cm height or less, depending on the level of decy. Chemicls were 193

7 dministered t 45 ngle (downwrd) to the erect portion of the decyed plnt, llowing introduction of the chemicl dose with no leching from the injection hole. Only the most recently hrvested residul llowing injection, where possible still ttched to the mother plnt, ws treted t ech mt. The lyout of the trils followed rndomised block design with three replictes. Plots hd pproximtely 50 plnts nd were seprted by two-row brrier. To stndrdise for biotic influences, replictes were orientted perpendiculr to the se/lnd breeze nd moisture grdient in the field. Appliction of pesticides (nd wter) ws conducted every 2nd month from lte October 2003 to lte August Yield, dmge prmeters nd pseudostem girth of plnts felled during 3-month period (August to October) in 2003 to 2005 were mesured. Yield ws determined t the pck-house by weighing of bnns (bunches excluding the peduncle). The plnts were subjected to weevil dmge nd girth smpling within week of hrvest. The Coefficient of Infesttion (CI) ws determined by pring the corm nd scoring the proportion of the rhizome circumference with weevil glleries (Vilrdebó 1973). Intervls of 2.5% were included up to level of 10% dmge. Dmge ws lso rted by the Percentge Coefficient of Infesttion (PCI) (Mitchell 1978, 1980), which involved scoring the presence/bsence of peripherl dmge for ten sections, ech covering 18 of the corm surfce. The ltter ws determined t 5 cm (Gold et l. 1994) nd between 5 nd 20 cm from the collr. The two PCI vlues were summed to provide totl PCI vlue. A cross section of the corm ws mde t 10 cm from the collr nd the percentge dmge of the centrl cylinder nd cortex scored in 10% intervls, using trnsprent circulr grid divided into 36 sections (modified from Gold et l. 1994, Kiggundu 2000). The two cross section vlues were verged to provide the men cross sectionl dmge (X men). The circumference of hrvested plnts ws mesured t 100 cm from the collr. Adult densities were ssessed over 4 weeks in October (from 2003 to 2005) nd April (from 2004 to 2005). Three split-pseudostem trps, plced individully next to three plnts in the middle of ech plot, were used to smple dult densities. Trp mteril ws rndomly selected from plnts hrvested within 2 weeks before trp preprtion t plnttion similr to, but seprted by dirt rod, from the specific tril sites. Only one trp ws prepred from ech plnt nd pseudostems with internl dmge/necrosis/tunnels were discrded. Pseudostem trps were 30 cm in length 194

8 (pseudostem section cm bove the collr), bisected longitudinlly nd ech hlf plced (with the cut surfce ventrlly) directly next to the mt of the plnt. Two hlves were plced on opposite sides of the mt nd regrded s one trp. The split pseudostems were covered with mulch to dely desicction nd decomposition. Trps were replced once week, when the smples per trp were counted nd destroyed. Nemtode smples were collected nd nlysed in October (from 2003 to 2005). Root smples were collected from three rndomly selected mother plnts per plot. Smples were sent to the ARC - Institute for Tropicl nd Subtropicl Crops (Nelspruit, Mpumlng), where 30 g of roots (rndomly selected per plot) were exmined for nemtodes. The initil dt were recorded before ny of the tretments were pplied Sttisticl nlysis Anlysis of covrince (ANCOVA) (Sokl nd Rohlf 1997) ws used to quntify yield nd girth over time, mong tretments nd between the interctions of time nd tretment. The nemtode number of ll the species were combined nd entered s covrite. Nemtode densities over time nd between tretments were scertined by fctoril ANOVA, while the pre-tretment densities were compred by one-wy ANOVA. The seven prmeters used for dmge estimtion were compred over time, between tretments nd mong the interctions of time nd tretment using repeted mesures ANOVA. One-wy ANOVA (Sokl nd Rohlf 1997) ws used to scertin pre-tril differences in dult densities. Differences of dult densities over time, between tretments nd between the interctions of the independent vribles were determined by fctoril ANOVA (Sokl nd Rohlf 1997). The Tukey HSD test (Sokl nd Rohlf 1997) ws used for ll post hoc nlysis. Unless stted otherwise, the dt were not trnsformed nd showed norml distribution nd homogeneity of vrinces in the liner scle. The STATISTICA Version 7 (Sttsoft Inc. 2004) softwre progrm ws used for nlysis. 6.3 Results Munster tril Yield, girth nd nemtodes No differences were found in yield (F 4, 9 = 0.29, P = 0.876) nd plnt girth (F 4, 9 = 195

9 1.21, P = 0.372) between plots t the onset (spring 2003) of the tril. The initil nemtode densities between plots were lso similr (F 4, 10 = 0.52, P = 0.724). Bunch yield ws sttisticlly similr in the spring of 2004 (24.80 ± (SE) kg) nd 2005 (26.59 ± (SE) kg) (F 1, 19 = 0.57, P = 0.460), ws no different between tretments (rnge: 24.9 to kg) (F 4, 19 = 0.38, P = 0.822) nd did not show significnt interction between time nd tretment (F 4, 19 = 0.57, P = 0.687). Similrly, plnt girth lso showed neither significnt temporl (2004: ± (SE) cm, 2005: ± (SE) cm) (F 1, 19 = 0.06, P = 0.813) or tretment effect (rnge: to cm) (F 4, 19 = 0.34, P = 0.845), nor n interction between the independent vribles (F 4, 19 = 1.43, P = 0.261). The nemtode-complex showed significnt difference between dtes (F 1, 20 = 5.97, P = 0.024), while numbers between tretments nd between interctions of time nd tretment were similr (F 4, 20 = 0.12, P = nd F 4, 20 = 0.45, P = 0.771, respectively). Post ANOVA nlysis showed tht the verge number of nemtodes ws significntly higher in 2005 (1770 nemtodes per 30 g roots) compred to 2004 (1000 nemtodes per 30 g roots). Anlysis of smples in 2003 nd 2004 showed tht the spirl (Helicotylenchus spp.) nd lesion nemtodes (Prtylenchus spp.) were pproximtely of equl proportions. Root smples in 2005 were minly infested with spirl nemtodes, but lesion nd root knot nemtodes (Meloidogyne spp.) were lso present. The burrowing nemtode, R. similis ws not found throughout the tril Dmge prmeters The pre-tril plnt dmge estimtions of the PCI (0-5 cm), PCI (5-20 cm), Totl PCI, CI, XO, XI nd X men were similr between plots (F 4, 10 = 0.81, P = 0.548; F 4, 10 = 0.18, P = 0.943; F 4, 10 = 0.39, P = 0.810; F 4, 10 = 0.75, P = 0.580; F 4, 10 = 0.77, P = 0.567; F 4, 10 = 0.46, P = nd F 4, 10 = 0.43, P = 0.787, respectively). The repeted mesures ANOVA showed no significnt difference between dte (2004 nd 2005) nd ny of the dependent vribles (0.06 < F 1, 20 < 4.05, 0.06 < P < 0.805). Significnt tretment effects were only found for PCI (5-20 cm) (F 4, 20 = 5.48, P = 0.004), Totl PCI (F 4, 20 = 5.56, P = 0.004) nd CI (F 4, 20 = 4.57, P = 0.009), but were minimised in the fipronil tretment for PCI (0-5 cm) (rnge: 0.63 to 1.94), XO (rnge: to 34.03%), XI (rnge: 2.31 to 15.44%) nd X men (rnge: to 23.89%). The nlysis showed n interction between dte nd tretment for X men (F 4, 20 = 3.46, P = 0.027). Post ANOVA nlysis found tht the PCI (5-20) 196

10 prmeter ws significntly lower in chlorpyrifos nd fipronil treted plnts compred to control plnts. Vlues for the remining chemicls were sttisticlly similr to ll the other tretments (Fig. 6.1). Totl PCI showed similr differences between mens s the PCI (5-20) dmge prmeter. Compred to the control, the coefficient of vrition (CI) ws only significntly lower in fipronil treted plnts (Fig. 6.1). The CI vlues for bifenthrin, chlorpyrifos nd oxmyl treted plnts were sttisticlly similr to the fipronil tretment (Fig 6.1). The men cross sectionl dmge ws significntly lower in plnts treted with fipronil in 2005 (7.94%) compred to control plnts in 2004 (32.78%) (dt not shown) Adult densities The one-wy ANOVA showed tht the pre-tril dult densities were similr between plots (F 4, 10 = 0.61, P = 0.666). The number of dults vried significntly over time (F 3, 40 = 3.33, P = 0.029) nd between tretments (F 4, 40 = 5.36, P = 0.001), but did not interct significntly between time nd tretment (F 12, 40 = 0.83, P = 0.621). Post ANOVA nlysis showed tht the men of 1.45 dults collected (in three trps per week) in October 2004 ws significntly less thn 3.72 dults collected in April Vlues of the other months were sttisticlly similr to October 2004 nd April 2005 (dt not shown). Fipronil treted plots hd significntly lower number of dults compred to the control (Fig. 6.2). The bifenthrin, chlorpyrifos nd oxmyl tretment were similr to the control nd fipronil treted plots (Fig. 6.2) Rmsgte tril Yield, girth nd nemtodes No differences were found in yield or plnt girth (F 5, 11 = 0.42, P = nd F 5, 11 = 1.34, P = 0.316, respectively) between plots t the onset of the tril. The initil nemtode densities between plots were lso similr (F 5, 12 = 2.06, P = 0.141). The yield in spring 2004 incresed from men of ± (SE) to ± (SE) kg per bunch in spring 2005, but the difference ws not significnt (F 1, 23 = 1.06, P = 0.314). There were no significnt differences (F 5, 23 = 0.55, P = 0.740) in yield between tretments, lthough the verge yield per bunch incresed by 11.29% in the fipronil (35.71 ± (SE) kg) nd 10.18% in the imidcloprid (35.27 ± (SE) kg) tretments compred to the control tretment (31.68 ± (SE) kg). The interction between time nd tretment lso showed no significnt 197

11 differences (F 5, 23 = 0.45, P = 0.808). Plnt girth ws similr between dtes (2004: ± (SE) cm, 2005: ± (SE) cm) (F 1, 23 = 1.71, P = 0.204). Plnt girth between tretments rnged from cm in the control to cm in the fipronil tretment (10.47% increse), but ws not significntly different (F 5, 23 = 1.16, P = 0.359). There ws no interction between time nd tretment (F 5, 23 = 0.43, P = 0.820). The density of nemtodes ws significntly different between dtes (F 1, 24 = 5.78, P = 0.024), but neither significnt tretment effect (F 5, 24 = 1.05, P = 0.414), nor n interction between time nd tretment ws found (F 5, 24 = 1.52, P = 0.221). The Tukey HSD test showed tht the nemtode number ws significntly higher in 2005 ( nemtodes per 30 g roots) compred to 2004 ( nemtodes per 30 g roots). Spirl nemtodes minly comprised the nemtode complex, but root knot nd lesion nemtodes were lso present t reltively low densities in 2003 to Burrowing nemtodes were present in two plots (verging 975 individuls per 30 g roots) during Dmge prmeters The pre-tril plnt dmge estimtions of PCI (0-5) (F 5, 12 = 2.59, P = 0.082), PCI (5-20) (F 5, 12 = 2.27, P = 0.113), CI (F 5, 12 = 0.93, P = 0.494), XO (F 5, 12 = 1.94, P = 0.161), XI (F 5, 12 = 1.46, P = 0.274) nd X men (F 5, 12 = 2.17, P = 0.126) showed no significnt difference between plots. The Totl PCI rnged from 9.44 (oxmyl tretment) to (fipronil tretment) nd ws different in the ANOVA nlysis (F 5, 12 = 3.66, P = 0.030), but not significntly differentited in the post hoc nlysis. The Tukey HSD test dopts conservtive pproch by employing experimentwise error rtes (bsed on the number of comprisons) for the type I error (Sokl & Rohlf 1997). The ANOVA showed no significnt difference between dte (2004 nd 2005) nd ny of the dependent vribles (0.01 < F 1, 24 < 0.43, < P < 0.939). Significnt tretment effects were found for ll the dmge estimtions; PCI (0-5) (F 5, 24 = 8.27, P < 0.001), PCI (5-20 cm) (F 5, 24 = 10.85, P < 0.001), Totl PCI (F 5, 24 = 13.98, P < 0.001), CI (F 5, 24 = 12.61, P < 0.001), XO (F 5, 24 = 9.38, P < 0.001), XI (F 5, 24 = 3.81, P = 0.011) nd X men (F 5, 24 = 9.40, P < 0.001). The nlysis showed n interction between dte nd tretment for PCI (0-5) (F 5, 24 = 2.95, P = 0.032). The Tukey HSD test showed the PCI (0-5) prmeter ws only significntly lower in fipronil nd imidcloprid treted plnts compred to control plnts. The oxmyl 198

12 tretment ws similr to the fipronil nd imidcloprid tretments, while bifenthrin ws similr to fipronil treted plnts (Fig. 6.3). PCI (5-20) showed tht only the fipronil nd imidcloprid tretments hd significntly lower dmge thn the control plnts, while the oxmyl tretment ws similr to fipronil nd imidcloprid. Reltive to the control plnts, the Totl PCI ws only significntly lower in oxmyl, fipronil nd imidcloprid treted plnts, while bifenthrin ws similr to the oxmyl tretment. All the chemicl tretments, except for chlorpyrifos, sttisticlly reduced the CI prmeter compred to untreted plnts; imidcloprid showed the lowest vlue which ws sttisticlly similr to fipronil nd oxmyl (Fig. 6.3). The XO ws only sttisticlly lower in the fipronil nd imidcloprid tretments compred to the control; oxmyl ws similr to fipronil nd imidcloprid treted plnts (Fig. 6.4). The imidcloprid treted plnts reduced the XI dmge by 100% nd ws the only tretment significntly lower thn the control (Fig. 6.4). Fipronil cused 95% reduction in XI. Imidcloprid, fipronil nd oxmyl significntly reduced the X men reltive to untreted plnts by 90.43, nd 63.48%, respectively. The bifenthrin tretment showed similr X men vlues to oxmyl nd fipronil (Fig. 6.4). The PCI (0-5) prmeter ws significntly lower in plnts treted with imidcloprid in 2005 (0.00) compred to control plnts in 2004 (1.83) nd 2005 (2.00) (dt not shown) Adult densities The distribution of dults between plots ws similr before the tril strted (F 5, 12 = 0.47, P = 0.794). Adult densities between subsequent collections rnged from 6.65 in April 2005 to 9.89 in October 2004 nd were significntly different (F 3, 48 = 2.99, P = 0.040). Tretment effects were lso significnt (F 5, 48 = 19.96, P < 0.001), but no interction between time nd tretment ws found (F 15, 48 = 1.08, P = 0.395). The Tukey HSD test did not show difference between collection dtes, becuse it dopts conservtive pproch by employing experimentwise error rtes (bsed on the number of comprisons) for the type I error (Sokl & Rohlf 1997). Fipronil, imidcloprid nd chlorpyrifos tretment resulted in significnt decrese in dult density compred to the control (Fig. 6.5). Bifenthrin nd oxmyl showed sttisticlly similr vlues to the control, while the chlorpyrifos tretment ws similr to the bifenthrin pplictions (Fig. 6.5). 199

13 6.4 Discussion Fipronil nd imidcloprid were highly effective chemicls ginst C. sordidus, minimising dmge to the periphery, cortex nd centrl cylinder of the rhizome nd significntly reduced dult density. The dmge prmeter of Cvendish bnns most closely relted to effective bunch weight is the percentge dmge to the centrl cylinder (Chpter 7). Fipronil nd imidcloprid virtully eliminted dmge to this portion of the rhizome fter six pplictions. Under certin conditions, the men percentge dmge to the cortex nd centrl cylinder of the corm cn be the best indictor of fruit yield (Chpter 7). This dmge ws lso gretly reduced fter six pplictions of fipronil nd imidcloprid. The percentge reduction in these importnt dmge prmeters fter chemicl ppliction should be considered s conservtive mesures. The mesurement scle of the cross sectionl dmge estimtes ws crude, with n increment (nd minimum) of 10% dmge. The result ws tht slight dmge (probbly less thn one percent) to some chemiclly treted plnts, especilly fipronil nd imidcloprid, ws scored s 10% dmge, while extensive lrvl tunnels comprising 10% re of the rhizome in control plnts received similr score. In future studies the estimte of percentge internl corm dmge should therefore be refined, preferbly to one percent scle. Of the other chemicl tretments, chlorpyrifos nd oxmyl showed reduction in peripherl dmge, but results were inconsistent. Injection of bifenthrin ws generlly ineffective. In previous study, single injection of chlorpyrifos during winter did not reduce tunnels in the rhizome of mother plnts fter 10 weeks (Dochez 1998). Bnn fruit yield nd plnt girth, corrected by nemtode densities, were not significntly incresed fter ny of the chemicl pplictions. Similr results hve been reported fter orgnophosphte nd crbmte tretment (Román et l. 1983; Chvrri-Crvjl & Irizrry 1997). Nevertheless, the dt showed (with nemtode infesttion constnt) n increse of up to 11.29% in effective bunch weight nd 10.47% increse in plnt girth in the fipronil nd imidcloprid tretments. This increse probbly would hve been significnt if the plot size ws incresed, thereby decresing the vribility. Moreover, plnts were propped during the tril nd plnt loss, which cn contribute more to yield loss thn reduction in bunch weight (Rukzmbug 1996), ws not considered. The Munster tril suggested tht (ssuming other vribles were constnt between the trils) if portion of residues injected with 200

14 fipronil over n nnul period is not ttched to the mother plnt, then n overll reduction in peripherl plnt dmge nd dult densities cn be expected, but reduction in the more importnt internl dmge estimtes my only be evident fter 2 yers. The results suggested tht fipronil nd imidcloprid, both considered to be systemic chemicls (Potter 1998; Nel et l. 2002), provided protective tretment when the injected residue is physiclly ttched to the mother plnt. The ppliction protocol used in this study is unique in tht systemic pesticides re injected throughout the yer. In Austrli plnts re lso injected with chemicls, but it is limited to contct pesticides pplied during winter (Treverrow pers. comm.). The high efficcy chieved fter injection of fipronil nd imidcloprid into plnt residues provides n optiml chemicl strtegy in n integrted pest mngement progrmme for the bnn weevil. Fipronil hs been shown not to ffect the vibility of Beuveri bssin Blsmo (Btist Filho et l. 1996). The pesticides belong to unique chemicl groups nd cn be sptilly nd temporlly ltered to minimise resistnce development. More importntly, the ppliction is specific to the pest, trgeting the residul plnt, which cn contin ll the weevil life stges throughout the yer, but where dults predominte (Chpter 3). These chemicls probbly lso provide control in the mother plnt, which cn contin ll the life stges, but where weevil lrve predominte (Chpter 3). The density of the nemtode-complex incresed during the tril nd ws not controlled by ny of the chemicl tretments. This is in generl greement with Pttison et l. (2002), who reported tht oxmyl injection into hrvested pseudostems ws not effective in controlling burrowing nemtodes. Injection of chlorpyrifos into post hrvest residues during winter lso provided no nemtode control (Dochez 1998). In ddition, no evidence of poisoning non-trget species ws observed during both field trils, lthough this spect ws not empiriclly evluted. In future, the ction mechnism nd residul ctivity of the pesticides under the ppliction protocol should be specificlly reserched. Timing of pplictions cn be optimised ccordingly, nd pplied when lrve (November to December, Februry) nd dults predominte (November, April/My nd July) (Chpter 3). 6.5 Acknowledgements Thnk you to Quentin Elliott nd Wyne Simpson for mking their frms vilble for 201

15 experimentl purposes nd providing ssistnce in the weighing of bnn bunches nd the smpling of dult beetles. Mieke Dneel (ARC-ITSC) nlysed the nemtode smples. BASF South Afric (Pty) Ltd., Byer (Pty) Ltd., Dow AgroSciences Southern Afric (Pty) Ltd., nd DuPont Nemours Int. S.A. kindly provided chemicls. The project ws finncilly supported by the Bnn Growers Assocition of South Afric (BGASA), Technology nd Humn Resources for Industry Progrmme (THRIP), Ntionl Reserch Foundtion (NRF) nd the University of Pretori (UP). 202

16 6.6 References ABERA, A.M.K., GOLD, C.S. & KYAMANYWA, S Timing nd distribution of ttck by the bnn weevil (Coleopter: Curculionide) in Est Africn highlnd bnn (Mus spp.). Florid Entomologist 82: ANITHA, N., CHARLES, J.S.K. & RAJAMONY, L Effect of sucker tretment with insecticides on incidence on bnn rhizome weevil (Cosmopolites sordidus). Indin Journl of Agriculturl Sciences 62: ANONYMOUS Beetle borer studies. Bnn Bulletin 33: 4. ANONYMOUS Agriculturl remedies. Ntionl Deprtment of Agriculture. World Wide Web: BATCHELDER, C.H Experimentos con insecticids pr combter el picudo negro, Cosmopolites sordidus (Germr), en plntciones de bc. Turrilb 4: BATISTA FILHO, A., LEITE, L.G., ALVES, E.B. & AGUIAR, J.C Control of Cosmopolites sordidus by fipronil nd its effect on Beuveri bssin. Arquivos do Instituto Biologico (So Pulo) 63: BRAITHWAITE, B.M Ground spryed tretments for control of the bnn beetle borer (Cosmopolites sordidus) (Germr). Journl of the Austrlin Institute for Agriculturl Science 24: BRAITHWAITE, B.M Bnn beetle borer control investigtions on the north cost of New South Wles. The Agriculturl Gzette of New South Wles 78: BUJULU, J., URONU, B. & CUMMING, C.N.C The control of bnn weevils nd prsitic nemtodes in Tnzni. Est Africn Agriculturl nd Forestry Journl 49: CÁRDENAS R El picudo negro del plátno Cosmopolites sordidus (Germr). Cenicfé 120: 12. CARDENAS MURILLO, R., ARROYAVE, R.F.P. & ARANGO BERNAL, L.G Tretment of plntin (Mus AAB Simmonds) ginst the bnn borer weevil (Cosmopolites sordidus Germr). Cenicfé 37: (Abstrct cited). CHABRIER, C., HUBERVIC, J. & QUENEHERVE, P Evlution of fosthizte (Nemthorin 10G) for the control of nemtodes in bnn fields in Mrtinique. Nemtropic 32:

17 CHAVARRIA-CARVAJAL, J.A. & IRIZARRY, H Rtes, ppliction intervls nd rottion of four grnulr pesticides to control nemtodes nd the corm-weevil (Cosmopolites sordidus Germr) in plntin. Journl of Agriculture of the University of Puerto Rico 81: COLLINS, P.J., TREVERROW, N.L. & LAMBKIN, T.M Orgnophosphorus insecticide resistnce nd its mngement in the bnn weevil borer, Cosmopolites sordidus (Germr) (Coleopter: Curculionide), in Austrli. Crop Protection 10: CUILLÉ, J Recherches sur le Chrncon du Bnnier. Institut des Fruits Agrumes Coloniux; Serie Technique No 4, Pris. DE JAGER, K., RABIE, E.C. & ANDERSON, S.H Chemiese beheer vn die piesngwortelboorder, Cosmopolites sordidus en lwurms, verl die boorlwurm, Rdopholus similis. Inligtingsbulletin - Nvorsingsinstituut vir Sitrus en Subtropiese Vrugte 228: DOCHEZ, C Study on Pest Sttus nd Control of Cosmopolites sordidus (Germr) in South Afric. M.Sc. Thesis, Heriot-Wtt University, Scotlnd. EDGE, V.E Cyclodiene-BHC resistnce in Cosmopolites sordidus (Germ.) (Coleopter: Curculionide) in New South Wles, Austrli. Bulletin of Entomologicl Reserch 64: 1-7. FOGAIN, R., MESSIAEN, S. & FOURÉ, E Studies on the bnn borer weevil in Cmeroon. Infomus 11: 8-9. FRANZMANN, B.A Bnn weevil borer in North Queenslnd. Queenslnd Agriculturl Journl 98: FROGGATT, J.L The bnn weevil borer (Cosmopolites sordidus). Queenslnd Agriculturl Journl 24: FROGGATT, J.L The Bnn Weevil Borer. Bulletin No. 4. Queenslnd Deprtment of Agriculture nd Stock, Brisbne, Austrli. GOLD, C.S., BAGABE, M.I. & SSENDEGE, R Bnn weevil (Cosmopolites sordidus) (Germr) (Coleopter: Curculionide) tests for suspected resistnce to crbofurn nd dieldrin in Msk district, Ugnd. Africn Entomology 7: GOLD, C.S. & MESSIAEN, S The Bnn Weevil, Cosmopolites sordidus. Mus Pest Fct Sheet No. 4, INIBAP. 204

18 GOLD, C.S., NIGHT, G., ABERA, A. & SPEIJER, P.R Hot-wter tretment for the control of the bnn weevil, Cosmopolites sordidus Germr (Coleopter: Curculionide), in Ugnd. Africn Entomology 6: GOLD, C.S., PEÑA, J.E. & KARAMURA, E.B Biology nd integrted pest mngement for the bnn weevil Cosmopolites sordidus (Germr) (Coleopter: Curculionide). Integrted Pest Mngement Reviews 6: GOLD, C.S., SPEIJER, P.R., KARAMURA, E.B., TUSHEMEREIRWE, W.K. & KASHAIJA, I.N Survey methodologies for bnn weevil nd nemtode dmge ssessment in Ugnd. Africn Crop Science Journl 2: JONES, R.K. & DIECKMANN, F Control of the bnn root borer, Cosmopolites sordidus Germr. Informtion Bulletin, Citrus nd Subtropicl Fruit Reserch Institute KIGGUNDU, A Host Plnt Rections nd Resistnce Mechnisms to Bnn Weevil, Cosmopolites sordidus (Germr) in Ugndn Mus Germplsm. M.Sc. Thesis, Ornge Free Stte University, South Afric. KOPPENHÖFER, A.M Observtion on egg lying behviour of the bnn weevil, Cosmopolites sordidus (Germr). Entomologi Experimentlis et Applict 68: MAOLIN, S Sttus of nemtode nd weevil borer problems ffecting bnn in Chin. In: Vlmyor, R.V., Dvide, R.G, Stnton, J.M., Treverrow, N.L., & Ro, V.N. (Eds.) Proceedings of Bnn Nemtode/Borer Weevil Conference in Los Bnos, Philippines, INIBAP MITCHELL, G The Estimtion of Bnn Borer Popultion nd Resistnce levels. Technicl Bulletin 2, Windwrd Islnd Bnn Growers Assocition (WINBAN), St. Luci. MITCHELL, G Bnn Entomology in the Windwrd Islnds. Finl report Centre for Overses Pest Reserch, London. NEL, A., KRAUSE, M. & KHELAWANLALL, N A Guide for the Control of Plnt Pests. 39th Edition. Directorte, Agriculturl Production Inputs, Deprtment of Agriculture, South Afric. NUNO, L. & RIBEIRO, V.P Cosmopolites sordidus in the utonomous region of Mdeir. Infomus 11:

19 PATTISON, A.B., VERSTEEG, C.A., MCQUINN, D.J., MATTHEWS, N.J., FARNSWORTH, W. & PIPER, R Mngement of burrowing nemtodes in bnns using pseudostem injection of Vydte L. DuPont, Austrlsi. PINESE, B. & ELDER, R Bnn Weevil Borer. Queenslnd Government, Deprtment of Primry Industries nd Fisheries, Horticulture nd Fresh Produce. World Wide Web: POTTER, D.A Destructive turfgrss pests. Ann Arbor Press, Michign. PRICE, N.S The use of modified pseudostem trpping technique for ssessing the efficcy of insecticides ginst the bnn borer weevil. Fruits 50: QUILICI, S Insect pests of bnn. Fruits 48: ROMÁN, J., ORAMAS, D. & GREEN, J Effect of rte nd frequency of ppliction of nemticides in plntin production. Nemtrópic 9: ROMÁN, J., ORAMAS, D., GREEN, J. & TORRES, A Control of nemtodes nd blck weevils in plntins. Journl of the Agriculturl University of Puerto Rico 67: RUKAZAMBUGA, N.D.T.M The Effects of Bnn Weevil (Cosmopolites sordidus Germr) on the Growth nd Productivity of Bnns (Mus AAA EA) nd the Influence of Host Vigour on Weevil Attck. Ph.D. Thesis, University of Reding, UK. SCHOEMAN, P.S Control of the bnn weevil Cosmopolites sordidus (Germr) (Coleopter: Curculionide). In: Robinson, J.C. (Ed.) Bnn Growers Assocition of South Afric 1: SCHOEMAN, P.S Chemicl control of the bnn root borer. Neltropik July: SCHOEMAN, P.S., SCHOEMAN, M.H. & DOCHEZ, C Novel techniques for the control of the bnn weevil, Cosmopolites sordidus (Coleopter: Curculionide) in South Afric. Yerbook of the Bnn Growers Assocition of South Afric 3: SHANAHAN, G.J. & GOODYER, G.J Dieldrin resistnce in Cosmopolites sordidus in New South Wles, Austrli. Journl of Economic Entomology 67: 446. SIMMONDS, N.W Bnns. 1st Edition. Longmns Press, London. SIMMONDS, N.W Bnns. 2nd Edition. Longmns Press, London. 206

20 SMITH, D Bnn weevil borer control in south-estern Queenslnd. Austrlin Journl of Experimentl Agriculture 35: SOKAL, R.R. & ROHLF, F.J Biometry. W.H. Freemn nd Compny, New York. STANTON, J.M Sttus of nemtode nd weevil borer problems ffecting bnn in Austrli. In: Vlmyor, R.V., Dvide, R.G, Stnton, J.M., Treverrow, N.L., & Ro, V.N. (Eds.) Proceedings of Bnn Nemtode/Borer Weevil Conference in Los Bnos, Philippines: INIBAP STATSOFT INCORPORATED Sttistic. Version 7. STOVER, R.H. & SIMMONDS, N.W Bnns. Third Edition John Wiley nd Sons, New York. TAYLOR, B Reserch field work on uplnd bnns, Mus spp., principlly cumint triploid AAA types, in the Knger region of Tnzni with observtions on growth nd on the cuses of decline in crop yield. Revist di Agricultur Subtropicle e Tropicle 85: TREVERROW, N.L Bnn Weevil Borer. Agfct H6.AE.1, 2nd edition. NSW Agriculture nd Fisheries, Austrli. TREVERROW, N., PEASLEY, D. & IRELAND, G Bnn Weevil Borer, A Pest Mngement Hndbook for Bnn Growers. Bnn Industry Committee, New South Wles Agriculture. VILARDEBÓ, A Resistnce of the bnn tree weevil Cosmopolites sordidus (Germr) to chlorinted hydrocrbon insecticides. Interntionl Congress of Plnt Protection 586. VILARDEBÓ, A Le coefficient d infesttion, critere d evlution du degre d ttques des bnneries pr Cosmopolites sordidus Germ. Le chrncon noir du bnnier. Fruits 28: VISWANATH, B.N Studies on the biology, vrietl response nd control of bnn rhizome weevil, Cosmopolites sordidus (Germr) (Coleopter: Curculionide). Mysore Journl of Agriculturl Sciences 11: VITTAYARUK, W., WATTANACHAIYINGCHAROEN, W., CHUAYCHAROEN, T. & WATTANACHAIYINGCHAROEN, D Sttus of weevil borer problems ffecting bnn in Thilnd. In: Vlmyor, R.V., Dvide, R.G, Stnton, J.M., Treverrow, N.L., & Ro, V.N. (Eds.) Proceedings of Bnn Nemtode/Borer Weevil Conference in, Los Bnos, Philippines: INIBAP

21 WILLERS, P., DANEEL, M.S. & DE JAGER, K Bnn. In: Vn den Berg, M.A., De Villiers, E.A. & Joubert, P.H. (Eds.) Pests nd Beneficil Arthropods of Tropicl nd Non-citrus Subtropicl Crops in South Afric. 36. ARC Institute for Tropicl nd Subtropicl Crops. Ad Dynmics, Nelspruit. WRIGHT, W.E Insecticides for the control of dieldrin resistnt bnn weevil borer Cosmopolites sordidus Germr. Austrlin Journl of Experimentl Agriculture nd Animl Husbndry 17:

22 Tble 6.1. Chemicl groups, trde nmes, formultions, ctive ingredients nd grm ctive ingredient of chemicls evluted ginst Cosmopolites sordidus t Rmsgte nd Munster (KZN, South Afric) from October 2003 to October * Excluded from the Munster tril. Chemicl group Trde nme (formultion) Active ingredient (.i.) Grm ctive ingredient (g..i.)/plnt Pyrethroid Tlstr (EC) Bifenthrin (100 g.l -1 ) Orgnophosphte Dursbn (WG) Chlorpyrifos (750 g.kg -1 ) Phenyl pyrzole Regent (SC) Fipronil (200 g.l -1 ) 0.01 Chloro-nicotinyl Confidor (SC) * Imidcloprid (350 g.l -1 ) Oxime crbmte Vydte (SL) Oxmyl (310 g.l -1 )

23 Figure legends Figure 6.1. The men vlues of the Percentge Coefficient of Infesttion (PCI) nd Coefficient of Infesttion (secondry xis) dmge prmeters of untreted (control) plnts nd plnts treted bimonthly with four chemicls t Munster (KZN, South Afric) from October 2003 to August For ech dependent vrible, mens with letters in common re not significntly different (P>0.05) nd upper cse letters refer to the secondry xis. 20 = PCI from > 5 to 20 cm from the collr, To = Summed totl PCI, Bifen = Bifenthrin, Chlor = Chlorpyrifos, Fip = Fipronil nd Ox = Oxmyl. Figure 6.2. The men dult bnn weevil density vlues of untreted (control) plots nd plots treted bimonthly with four chemicls t Munster (KZN, South Afric) from October 2003 to August For ech dependent vrible, mens with letters in common re not significntly different (P>0.05). Bifen = Bifenthrin, Chlor = Chlorpyrifos, Fip = Fipronil nd Ox = Oxmyl. Figure 6.3. The men vlues of the Percentge Coefficient of Infesttion (PCI) nd Coefficient of Infesttion (secondry xis) dmge prmeters of untreted (control) plnts nd plnts treted bimonthly with five chemicls t Rmsgte (KZN, South Afric) from October 2003 to August For ech dependent vrible, mens with letters in common re not significntly different (P>0.05) nd upper cse letters refer to the secondry xis. 05 = PCI from 0 to 5 cm from the collr, 20 = PCI from > 5 to 20 cm from the collr, To = Summed totl PCI, Bifen = Bifenthrin, Chlor = Chlorpyrifos, Fip = Fipronil, Imi = Imidcloprid nd Ox = Oxmyl. Figure 6.4. The men vlues of the cross sectionl dmge prmeters of untreted (control) plnts nd plnts treted bimonthly with five chemicls t Rmsgte (KZN, South Afric) from October 2003 to August For ech dependent vrible, mens with letters in common re not significntly different (P>0.05). XO = Cross section dmge percentge of the cortex, XI = Cross section dmge percentge of the centrl cylinder, X men = Averge cross sectionl dmge of the corm, Bifen = Bifenthrin, Chlor = Chlorpyrifos, Fip = Fipronil, Imi = Imidcloprid nd Ox = Oxmyl. 210

24 Figure 6.5. The men dult bnn weevil density vlues of untreted (control) plots nd plots treted bimonthly with five chemicls t Rmsgte (KZN, South Afric) from October 2003 to August For ech dependent vrible, mens with letters in common re not significntly different (P>0.05). Bifen = Bifenthrin, Chlor = Chlorpyrifos, Fip = Fipronil, Imi = Imidcloprid nd Ox = Oxmyl. 211

25 Figure 6.1 Men percentge (+SE) B AB b AB AB b b b b A b 20 To 20 To 20 To 20 To 20 To Coefficient of Infesttion (±SE) Bifen Chlor Fip Ox Control Percentge Coefficient of infesttion clss nd tretment 212

26 Figure 6.2 Men dult density (+SE) b b b b Bifen Chlor Fip Ox Control Tretment 213

27 Figure 6.3 Men percentge (+SE) A AB b BC b BCD bc CD b c D c b c bc bc b b To To To To To To Bifen Chlor Fip Imi Ox Control Percentge Coefficient of Infesttion clss nd tretment Coefficient of Infesttion (±SE) 214

28 Figure Men percentge (+SE) b b b bc b c bc b b bc c c b b XO XI X m XO XI X m XO XI X m XO XI X m XO XI X m XO XI X m Bifen Chlor Fip Imi Ox Control Cross section dmge clss nd tretment 215

29 Figure Men dult density (+SE) bc c b b 0.0 Bifen Chlor Fip Imi Ox Control Tretment 216