Journal of the American Mosquito Control Association, ll(2):260-269, lgg5Copyright @ 1995 by the American Mosquito Control Association, Inc.

INTRODUCTION

The use of synthetic chemical insecticides forvector control is in decline due to high costs, thedevelopment of resistance in many target pop-ulations, and perceived risks to the environmentand human health. Althoueh chemical insecti-cides will remain important in vector controlprograms, problems they have caused, and thepaucity ofnew types under development, havefor some time stimulated interest in alternativecontrol agents. In seeking replacements forchemicals, the pathogens and nematodes thatcause diseases fatal to vectors have received con-siderable study over the past 30 years, particu-larly with respectto theirpotential for controllingthe mosquito vectors of malaria and filariasis,and the blackfly vectors ofonchocerciasis (Fed-erici 1981, Chapman 1985, Lacey and lJndeen

THE FUTURE OF MICROBIAL INSECTICIDES ASVECTOR CONTROL AGENTS'

BRIAN A. FEDERICI

Department of Entomology and Graduate Program in Genetics,University of California, Riverside, CA 92521

ABSTRACT. Insect vectors of human diseases are subject to diseases of their own caused by viruses,bacteria, fungi, protozoans, and nematodes. Over the past 30 years, many members ofthese groups havebeen evaluated as vector control agents, particularly for mosquito control. Most pathogens and nematodesoccur primarily in larvae, and are only effective against this stage. The principal candidate control agentsstudied include iridescent and nuclear polyhedrosis viruses, thebacteia Bacillus thuringiensis and Bacillussphaericus, the fungi Lagenidium giganteum, Culicinomyces clavosporus, and species ofthe genus Coe-lomomyces, the protozoan Nosema algerae, and the mermithid nematode Romanomermis culicivorax.Of these, the only one considered an operational success is the bacterium, Bacillus thuringiensis subsp.israelensis (B.t.t.), which has proven usefirl for control of both mosquito and blackfly larvae in programswhere larviciding has been traditionally employed as a vector control tactic. The reasons for the successof B.t.i. are its cost-efectiveness and relative ease of use, which are due, respectively, to the ability ofB.t.i. to be grown on artificial media and the development of formulations that can be applied usingconventional insecticide application technology. Because few microbial insecticides are cost-effective, andthose that are are only effective against larvae, these agents will likely play only a minor, but in somecases irnportant, role in most future vector control programs.

t This paper was prepared for oral presentation aspart of a symposium, and thus my treatment of thesubject has been necessarily brief, particularly with re-spect to the data that serve as the basis for my opinicns"For those readers interested in a more detailed treat-ment ofeach ofthe groups I dealt with, and referencesto numerous original papers, I recommend the follow-ing reviews. For viruses, Federici (1974); for bacteria,Guillet et al. (1990) and Mulla (1990); for fungi, Fed-erici (1981), Hall and Papierok (1982), and McCoy etal. (1988); for protozoa, Henry (1981) and Brooks(1988); and for nematodes, Petersen (1982). The re-view on microbial control of mosquitoes and blackfliesby Lacey and Undeen (1986) is also very worthwhile,as is the book on bacterial control of mosquitoes andblackflies edited by de Barjac and Sutherland (1990).

1986, Guillet et al. 1990, Mulla 1990). Thesestudies have included the search for pathogensand nematodes, the laboratory and field evalu-ation of those that appeared to have the bestpotential for operational use, the developmentand assessment of methods for mass production,and finally implementation in operational con-trol programs followed by evaluation of cost-effectiveness based on control ofthe target pestsor vectors. Overall, these studies have resultedin 3 principal findings: l) pathogens and nema-todes are only effective against the larval stagesofvectors, 2) effective control typically requiresrepeated rather than a single application of theagent during the breeding season, and 3) to beused cost-effectively in vector control programs,a method for mass production of the control agentin vitro must exist. The first finding is simply abiological property of most of the pathogens andnematodes discovered to date; they may occurin adult stages, but seem to take their greatesttoll against larvae. The second results from nat-ural diminution, or what might be consideredpoor recycling potential from the standpoint ofvector control. The third results from the rela-tively large areas that must be treated, and de-rives from the high costs of mass rearing controlagents in organisms, such as mosquito or blackflylarvae, in comparison to the lower costs of massproduction for a control agent grown on an in-expensive artificial medium. The latter is onlypossible at present for some of the bacteria andfungi. As a result of the combination of thesefindings, only the bacterium, Bacillus thurin'giensis sttbsp. israelensis (B.t.i.), has been used

260

JuNe 1995 VBcron CoNrnol wrrHoLIT Cnwnclrs 261

in operational vector control programs, and onlyin programs where larviciding has been a tradi-tional method of vector control. Progress in de-veloping microbials for vector control has beenslow, and this situation is unlikely to improvesoon. Thus, it can be concluded that bacteria arelikely to be the only microbial agents used forsome time to come in vector control programs,and these will only be used in programs wherelarviciding has been a successful vector controlstrategy.

Predictions about the future use of microbialinsecticides and nematodes in vector controlprogxams are best arrived at by an examinationofwhat we have learned over the past 30 yearsfrom the study ofthese potential control agents.Thus, in this paper I will review the key findingsfor each of the major pathogen groups and nem-atodes that led to our current view oftheir prob-able use in the future.

VIRUSES

All viruses are obligate intracellular parasites,and as such must be grown in living hosts. Withinsect viruses, this means either in insects (i.e.,in vivo) or in cultured insect cells (in vitro) be-cause no methods exist for growing viruses on"artificial" media.

Despite these difficulties, searches were con-ducted for viruses pathogenic to mosquitoes andblackflies because at the time this research wasinitiated in the 1960s, it was thought that virusesmight work as classical biological control agents(i.e., agents that would lead to long-term controlonce introduced against previously unexposedpopulations), or that it might be possible to de-velop suitable methods for the mass productionof viruses using colonized insects. The search forviruses, carried out primarily by members of theUSDA's mosquito research laboratories at LakeCharles, LA, and Gainesville, FL, resulted in thediscovery ofnumerous viral pathogens oflarvae(Federici 1974). These included several irido-viruses of floodwater mosquitoes belonging tothe genera Aedes and Psorophora, cytoplasmicpolyhedrosis viruses (CPVs), and baculovirusesof the nuclear polyhedrosis virus (NPVs) type,the latter 2 types being isolated primarily fromAedes and Culex species. Iridoviruses and CpVshave also been reported from larvae of severalblackfly species.

Studies of the iridoviruses. or iridescent vi-ruses as they are more commonly known, showedthat they infected and replicated in most larvaltissues with the exception of the midgut epithe-lium (Anthony and Comps l99l). However, inIaboratory trials these viruses were shown to beonly moderately infectious for larvae, with mor-

tality rates typically being less Ihan 25o/o for I stand 2nd instars, even when larvae were placedin water containing virions at concentrations ofseveral g.g/ml. The NPVs and CPVs were shownto only infect the midgut epithelium, where theNPVs replicated in the nuclei throughout the gut,and the CPVs in the cytoplasm of infected cellsin the gastric ceca and posterior stomach. As withthe iridescent viruses, the mortality rates ob-tained with NPVs rarely exceeded 25o/o, and al-though higher rates of infection were achievedwith CPVs, even larvae heavily infected with thisvirus type could often survive the infection, pu-pate, and emerge as adults. Owing to their poorefrcacy and lack of suitable methods for massproduction, serious research on the developmentand evaluation of viruses as mosquito or blackflycontrol agents was discontinued in the 1970s.

BACTERIA

Bacteria are relatively simple unicellular mi-croorga.nisms that lack internal organelles suchas a nucleus anC mitochondria, and reproduceby binary fission. With few exceptions, most ofthose discovered in vectors grow readily on awide variety of inexpensive substrates or artifi-cial media, a characteristic that greatly facilitatestheir mass production. The bacteria currentlyused or under development as microbial controlagents for vectors are spore-forming members ofthe bacterial family Bacillaceae. Two species havereceived considerable stldy, Bacillus thuringien-srs subsp. israelensis (B.t.i.), originally discov-ered in Israel by Goldberg and Margalit (1977)and the only species currently used operationally,and Bacillus sphaericus, which has good larvi-cidal activity against certain mosquito larvae,but whose operational utility has not been clearlydemonstrated.

Both B.r.i. and B. sphaeriao kill larvae throughthe action of insecticidal proteins that destroythe larva's midgut epithelium shortly after in-gestion, usually within a few hours. These insec-ticidal proteins are contained within a parasporalbody produced when the bacterium sporulates(Fig. l). In B.t.i.,1he parasporal body is spherical,and contains 4 proteins with molecular massesof 27, 7 2, I 28, and I 34 kDa (Federici et al. I 990).These proteins interact synergistically to yieldtoxicity comparable to that of chemical insecti-cides such as DDT. Bacillus thuringiensis israe-lensis h;as been shown to be toxic to most dip-terans of the suborder Nematocera, includingmosquitoes (Culicidae), blackflies (Simuliidae),craneflies (Tipulidae), and midges (Chironomi-dae). One ofthe advantageous properties of.B.r.i.is that it is highly insecticidal for many vectorspecies, including species of Aedes, Culex,

Juxe 1995 VEcroR CoNTRor wITHour Cn*ncers 263

Table L Comparison of the efrciency of production relative to application rates for -representative bacterial and fungal pathogens evaluated as microbial insecticides for control of

mosquito larvae.

Agent LCrolhectareProduction

yield Medium/hectaret

BacteriaB aci I lus t huri n giens is

subsp. israelensisBacillus sphaericus 2362

FungrC ulicino myces c I av o sp orusMaarhizium anisopliae

200 g technicalpowder2

100 g technicalpowder2

l0t'conidia2 kg conidia

l0 g/liter

10 g/liter

20 liters

l0 liters

l0rt conidia/liter 1,000 litersI kg25 kg rice 50 kg ofrice

t Refers to the amount of liquid or solid substrate needed to produce the amount of technical matcrial required to achievepopulation reductions of 90% for I ha.

, Technical powder rcfers to the unformulated dried solids obtahed after fermentation. Of this weight, only approxinarely25% consists of insecticidal proteins in B. thuringiensis, and 5-10% in B. sphaerias; the rcst is fermentation solids, bacterialspores, and rcmnants ofbacterial cells.

B.t.i. used in vector control programs are Acrobe(American Cyanamid), Skeetal (Novo Nordisk),Teknar (Sandoz), and Vectobac (Abbott l^abo-ratories). Although not used widely for vectorcontrol, these products have proven useful andcost-effective in programs that employed larvi-cides as a vector control tactic. In addition, theyare commonly used in industrialized countriesfor the control ofnuisance mosquitoes and black-flies, particularly in environmentally sensitiveareas.

With respect to B. sphaericas, there are 3 in-secticidal proteins, with molecular masses of 43,52, and 100 kDa, but these do not alwavs occurin all isolates @aumann et al. I 991, Porter et al.

I 993). Interestingly, these proteins are only toxicto mosquitoes, and tend to exhibit the highesttoxicity against Culex species. No significant tox-icity for B. sphaericus has been reported againstblackflies or other dipterans. Though the spec-trum of activity for 8. sphaericus is not as broadas for B.t.i., B. sphaericus has greater residualactivity, and works better in highly polluted wa-ters, although it requires high rates of applica-tion, generally greater than I g oftechnical pow-der/m2.It is still considered to hold potential forthe species of Culex that vector filarial worrns(Hougard 1990, Yap 1990), however, no com-mercial products based on B. sphaericus are cur-rently marketed. This does not mean that for-

Table 2. Summary of property ratings for selected pathogens and nematodes evaluated forveclor controt.

Bacteria Fungus NematodeVirus

Property

Iridovirusor

baculovirus

Bacillusthuringiensis

subsp.israelensis

Culici-nomycesclavo-sporus

Protozoan

Nosemaalgerae

Romano-mermiscalici-vorax

Broad host/targetspectrum

Mass productionIn vitroIn yiyo

Efficacy/unit weightSafety to nontargetsFormulationStorageResidual activity

+ +

N/P'+ ++ ++ + + + + ++ + + + ++ + ++

+ + + + +

+ + + + + +N/A2+ + + + + ++ + + + ++ + + + + ++ + + + + ++

+ + + + + + + + + + +

+++ N/P N/P+ + + + + ++ + + + + ++ + + + + + + + + + + ++ + + + + + + + + ++ + + + + + + ++ + + + + +

I N/P : Not practical or cost-effective with existing technology.'� N/A : Not applicable because technology already exists foicost-effective mass prod.uction in vilro.

264 JounNer op rxs Ar,tsRrcAN Mosqurro CoNrnol AssocrarroN VoL 11, No. 2

mulations based on B. sphaericus would not beuseful in filariasis control programs, but ratherthat the market is considered too small to meritcommercial development until more cost-effec-tive formulations are devised.

FUNGI

Three types of aquatic fungi that attack mos-quito larvae have been studied for use as bio-logical control agents, species of Coelomomyces(chytridiomycete fungi), Lage nidium giganteum(an oomycete fungus), and Culicinomyces cla-vosporus (a deuteromycete fungus).

The genus Coelomomyces comprises more than70 species of obligately parasitic fungi that un-dergo a complex life cycle involving an alterationof sexual (gametophytic) and asexual (sporo-phytic) generations (Couch and Bland 1985.Whisler 1985). In all species studied, the sexualphase parasitizes a microcrustacean host, usuallya copepod, whereas the asexual generation typ-ically develops in a mosquito larva. In the lifecycle, a biflagellate zygospore invades the hem-ocoel of a mosquito larva where it produces asporophyte that colonizes the body and formsresistant sporangia. The larva dies and subse-quently the sporangia undergo meiosis, produc-ing uniflagellate meiospores that invade the hem-ocoel of a copepod host where a gametophytedevelops. At maturation, the gSmetophyte cleavesforming thousands of uniflagellate gametes.Cleavage results in death of the copepod andescape of the gametes, which fuse forming bi-flagellate zygospores that seek out another mo$-quito host, completing the life cycle. The lifecycles ofthese fungi are highly adapted to thoseof their hosts. Moreover, as obligate parasitesthese fungi are very fastidious in their nutritionalrequirements, and, as a result, no species of Coe-lomomyces has been cultured in vitro.

Coelomomyces is the largest genus of insect-parasitic fungi, and has been reported worldwidefrom numerous mosquito species, many ofwhichare vectors of important diseases such as malariaand filariasis. In some of these species, for ex-ample, Anopheles gambiae Giles in Africa, epi-zootics caused by Coelomomyces kill more than950/o of larval populations in some areas (Couchand Umphlett 1963). Such epizootics led to ef-forts to develop several species as biological con-trol agents over the past 30 years (Federici I 98 I ).However, these efforts have largely been discon-tinued due to the discovery that the life cyclerequired a 2nd host for completion, the inabilityto culture these fungi in vitro, and the develop-ment of B.t.i. as a bacterial larvicide for mos-quitoes.

Although the difrculties encountered withCoelomomyces make it unlikely that this funguswill ever be developed as a biological controlagent, there is still considerable interest in La-genidium giganteum. This oomycete fungus has2 important advantages over Coelomomyces; itis easily cultured on artificial media and it doesnot require an alternate host (Federici 1981, Ja-ronski et al. 1983). In the life cycle, a motilezoospore invades a mosquito larva through thecuticle. Once within the hemocoel, the funguscolonizes the body over a period of 2-3 days,producing an extensive mycelium consistinglargely ofnonseptate hyphae. Toward the end ofgowth, the hyphae become septate, and out ofeach segment an exit tube forms that grows backout through the cuticle and forms zoosporangiaat the tip. Zoospores quickly differentiate in these,exiting out through an apical pore to seek out anew substrate. In addition to this asexual cycle,thick-walled resistant sexual oospores can alsobe formed within the mosquito cadaver.

Techniques have been developed to produceboth zoosporangia and oospores in vitro, andmethods are currently being developed to modifyexisting technology so that the fungus can be massproduced. Several years offield trials in Califor-nia and North Carolina have shown that the zo-osporangia are too fragile for routine use in op-erational control programs (Merriam and Axtell1982, Kerwin and Washino 1987). The oospore,however, is quite stable, although germinationremains unpredictable. Nevertheless, freld re-sults indicate that germination of even a smallpercentage ofoospores can result in the initiationof epizootics that lead to season-long mosquitocontrol. The fungus is, however, quite sensitiveto ammonia and chloride ions, and thus is notefrcaceous in polluted or brackish waters. Nev-ertheless, although several technical problems re-lated to mass production and formulation re-main to be overcome, L. giganteurn is a prom-ising candidate for successful commercial de-velopment and use in freshwater habitats suchas rice fields. Its principal advantage over B.t.i.is that if efective formulations of the oosporecan be developed, it appears that in many hab-itats only a single application would be requiredper season. Even less frequent application maybe possible in some habitats, as evidence suggeststhat the oospores can overwinter, initiating ep-izootics the following seasons. The extent to whichthis occurs and can be relied upon for effectivemosquito control remains to be determined.Large-scale field trials have not yet been con-ducted with L giganteum, so production costsand rates ofapplication cannot be related to thelevel and length ofvector control. Thus, it is notpossible to assess the cost-effectiveness of L. gi-

Vscror Coxrnor, wrrHour Cnnacets 265JUNB 1995

ganteum as a mosquito or vector control agentat this time.

In addition to Coelomomyces and L. gigan-teum, the aquatic deuteromycete fungi, Culici-nomyces clavosporw and Tolypocladium cylin-drosporum, and the terrestrial deuteromycetefungus, Metarhizium anisopliae, have been con-sidered for mosquito control (Federici 1981,Sweeney et al. 1983, Soares and Pinnock 1984).These fungi produce conidia, which adhere tothe cuticle of larvae, and then invade and colo-nize the larva via a germ tube. However, highproduction costs resulting from inefrcient pro-duction methods. lack of cost-effective controlin the field, and the discovery and developmenlof B.t.i., have eliminated these fungi as seriouscandidates for development as microbial controlagents.

PROTOZOA"Protozoa" is a general term applied to a large

and diverse group of eucaryotic unicellular mo-tile microorganisms that belong to what is nowknown as the kingdom Protista. Members of thiskingdom can be free-living, saprophytic, com-mensal, symbiotic, or parasitic. The cell containsa variety of organelles, but has no cell wall, andcells vary greatly in size and shape among dif-ferent species. Feeding is by ingestion or moretypically by absorption, and vegetative repro-duction is by binary or multiple fission. Bothasexual and sexual reproduction occur and thelatter can be very complex, and is often usefulfor taxonomic purposes. Many protozoa producea resistant spore stage that is also used in tax-onomy. The kingdom is divided into a series ofphyla based primarily on the mode of locomo-tion and structure of locomotory organelles, andincludes the Sarcomastigophora (flagellates andamoebae), Apicomplexa (sporozoa), Microspora(microsporidia), Acetospora (haplosporidia), andCiliophora (ciliates). Some types ofprotozoa, suchas the free-living amoebae and ciliates, are easilycultured in vitro, whereas most of the obligateintracellular parasites have not yet been grownoutside ofcells.

As might be expected from such a large anddiverse group of organisms, many protozoans areparasites of vectors. Those that have been re-ported most commonly in vector populations aremicrosporidia (phylum Microspora), and thus itis these, especially members of the genera No-sema and. Amblyospora. that have received themost study. These studies have shown that mostmicrosporidians have the general feature ofcaus-ing chronic diseases. Thus, most cannot be usedas fast-acting microbial insecticides. This char-acteristic is not detrimental in the case of most

vectors because it is the adults that are respon-sible for disease transmission. However, the mi-crosporidia are all obligate intracellular para-sites, and there is no easy or cheap way to massproduce them in vitro. M'oreover, the few fieldtrials that have been conducted indicate that veryhigh rates of application are required to obtainsignificant reductions in vector populations, atleast in mosquito populations, the only vectorsthey have been tested against. The lack of com-mercially suitable methods for mass productionand poor efrcacy have significantly diminishecinterest in developing protozoans as vector con-trol agents. Fairly extensive studies were carriedout with Nosema algeraein the 1970s to evaluateits potential for controlling anophelines, with theaim of reducing the prevalence of malaria, andthus these studies are worth reviewing brieflyhere.

Nosema algerae was originally isolated fromAnopheles stephensiListon, but has a broad hostrange against anophelines and is capable underexperimental conditions ofinfecting a wide rangeof other mosquito species. It is also capable ofreproduction in the lepidopteran Heliothis zea,and this host has been used to produce sporesfor field trials (Anthony et al. 1978b). The in-fectious stage is a spore, within which is a coiledtube, the polar filament, that is responsible forinjecting the sporoplasm into larvae after inges-tion. Once within the body of alawa,the Nosemacells undergo numerous merogonic cycles in lar-val tissues, and then a sporogonic cycle that re-sults in the development of new spores. The lar-vae typically die at the end ofthe disease iftheyare infected during the early instars. Ifinfectedduring the 4th instar, larvae will often pupateand adults emerge. These adults have lower fe-cundity and decreased capacity for malaria trans-mission (Anthony et al. 1978a). Although thelatter results initially indicated some potentialfor having an impact on disease transmission,field trials carried out in Panama against Anoph-eles albimanus Wied. in the 1970s showed thatvery high and uneconomic application rates forspores would be required to have any significantimpact on disease control. To achieve mosquitoinfection rates of 860/o, 4 applications of sporesat a rate of 2.2 x lOe/m2 had to be made. Totreat a hectare at such a rate would require aminimum of approximately 40,000 lepidopteranlarvae to produce the needed amount ofspores.

Members of the genus Amblyspora also occurcommonly in mosquito populations and havereceived considerable study. Studies of severalspecies of Amblyospora have shown that theyhave complex life cycles, much like those of Coe-lomomyces, requiring an alternate crustacean host(Sweeney et al. 1985). As a result, these parasites

266 Jounrqer or rxr ArraunrceN Mosquro Covrnol AssocnrroN Vou 11 , No .2

are also thought to have little potential for de-velopment as vector control agents.

MERMITHID NEMATODES

Mermithid nematodes are among the largestnematodes attacking insects, and the adult fe-males typically measure from 5 to over 20 cm.In parasitized insects with a translucent cuticle,such as the larvae of mosquitoes and blackflies,the advanced stages ofdeveloping nematodes canoften be observed within the hemocoel wherethey appear as long, thin white worrns. Mermi-thids are obligate parasites and have been re-ported from many different orders of insects aswell as from other arthropods such as spidersand crustaceans. However, the only ones seri-ously considered for use as biological controlagents are the species Romanomermis culicivo-rax and Romanomermis iyengari, which are ca-pable of parasitizing many species of mosquitolarvae (Petersen 1982). The life cycle of R. cul-icivorax, which is quite typical of mermithids,will be used to illustrate the properties that re-sulted in studies ofthe biological control poten-tial of this nematode type.

The females of R. culicivorax are found in wetsoil at the bottom of aquatic habitats in whichmosquitoes breed. Here, after mating, the fe-males lay thousands of eggs. The embryo devel-ops into a lst-instar juvenile over a period ofabout I wk, and afterwards molts to a 2nd-stagejuvenile still within the egg. This 2nd stage (pre-parasite) then hatches out and swims to the watersurface where it seeks out and, with the aid ofthe stylet, invades an early instar mosquito larvavia the cuticle. Once within the hemocoel, theimmature larva grows over a period of 7-10 daysby absorbing nutrients through its cuticle, andmolts once during this period. Upon completionof this parasitic phase, the 3rd-stage juvenilepunctures its way out through the cuticle of itshost, thereby killing the mosquito, and then dropsto the bottom ofthe pond where it matures with-out feeding over another 7-10 days, finally molt-ing to the adult stage. The adults then mate andthe females lay eggs, thus completing the life cy-cle.

Using several different species of colonizedmosquitoes and simple rearing techniques, Pe-tersen (1982) and his colleagues were able to de-velop mass rearing techniques for R. culicivorax,and conducted a wide range of trials in whichthey evaluated this nematode by applying thepreparasitic stage in the laboratory and field. Inthe field studies, rates of parasitism (: mortality)for anopheline larvae as high as 850/o were ob-tained when preparasites were applied at a rateof 2,400/m2. Even higher levels of parasitism

(960/o) were reported by kvy and Miller (1977)when they applied preparasites at a rate of 3,600/m2 against floodwater mosquitoes in Florida.When introduced into suitable habitats, R. cul-icivorax can recycle, providing some level of con-trol in subsequent generations ofmosquitoes, butthis level is typically not sufficient to interruptthe vector potential of mosquitoes or alleviatethe annoyance problem. In addition, both lab-oratory and field studies showed that R. culici-vorex was quite sensitive to chloride ions, andthe methods developed for mass production,storage, shipment, and use were found not to becost-effective, particularly in comparison tocommercial formulations of the bacterial insec-ticide B. thuringiensis subsp. rsralensls, whichwas discovered while the development and eval-uation of R. culicivorexwas underway. As a re-sult of its limitations, and the availability ofcheaper and more effective alternative biologicalmethods of control, there is little interest todayin the further development of R. culicivorax, orother mermithids for that matter, as biologicalcontrol agents.

SUMMARY ANDCONCLUSIONS

During the past 3 decades, in response to theproblems caused by synthetic chemical insecti-cides, a considerable effort has been made to findand develop pathogens and nematodes as bio-logical control agents, and microbial insecticidesfor control of vectors. This research has pro-duced a wealth ofknowledge about the occur-rence, basic biology, and control potential ofnu-merous viruses, bacteria, fungi, protozoa, andnematodes that parasitize vector insects. Thisknowledge is of value in its own right and servesas a foundation against which to measure thepotential ofnew control agents. However, studiesof the various organisms that have been discov-ered and evaluated as control agents, in combi-nation with an assessment of how they might beused in vector control programs, have demon-strated that most of these agents are not suffi-ciently cost-effective to be used in such controlprograms. Moreover, there is no reason to be-lieve this assessment will change over the nexl10-20 years, barring some sort ofunanticipatedbreakthrough.

The primary reason for the lack of cost-effec-tivenes$ for most pathogens and nematodes thathave been evaluated is that methods are notavailable for their mass production at an ac-ceptable cost. In essence, to achieve levels ofvector control that will have a significant impacton the prevalence of human disease, the amountsof vector control agents that would have to be

JUNE 1995 VBcron CoNrnol wITHour CnslurcAls 267

applied are too high, and thus too costly to pro-duce, even where labor costs are low, using thein vivo melhods that are required to produce theseagents. Moreover, even when applied at highrates, none ofthe pathogens or nematodes eval-uated persist or recycle at a sufficiently high levelto give acceptable rates ofvector control for morethan I month. Despite the numerous potentialcontrol agents evaluated, the only cost-effectivepathogen identified to date is the bacterium Ba-cillus thuringien.ris subsp. israelensis. Its successis due primarily to its ability to be mass producedin large quantities on relatively inexpensive me-dia using large-scale fermentation technology.Even so, this bacterium is only effective againstlarvae, and thus its use is limited to vector con-trol programs that have traditionally used larv-iciding as a tactic. Though limited, the use ofB./.1. has been quite valuable in The World HealthOrganization's Onchocerciasis Control Programin West Africa, and may ultimately prove usefulin malaria, filariasis, and viral disease controlprogams that include laryiciding. However, B.t.i.is unlikely in the future to be of use against majormalaria vectoni such as An. gambiae and, An.stephensi, where larviciding is not an importantcontrol strategy.

The lack of new agents and technologies tocontrol vectors indicate that chemical insecti-cides will continue to be important, where resis-tance is not a major problem, in vector controlprograms. New disease control technologies arebeing explored, such as disease-refractory trans-genic mosquitoes (Collins and Besansky 1994)and releasing transgenic insecticidal organismsused as larval food into breeding habitats. Theformer tactic has merit, but may ultimately leadto resistance in the population of the diseasecausing organism, such as the malarial parasites.The latter tactic, based on our experience withthe development of resistance to chemical in-secticides in larval populations, will almost cer-tainly lead to resistance.

In summary, the insect populations that vectorthe major diseases of humans, especially giventhe increasing human populations in areas wherethese diseases are the most prevalent, offer vectorcontrol challenges greater than ever. pathogens,especially bacteria such as B.t.i., wrll be of someuse in vector control programs, but will not pro-vide a major vector control tactic in the foresee-able future.

REFERENCES CITEDAnthony, D, W. and M. Comps. 1991. Iridoviridae,

pp. 55-86. In: J. R. Adams and J. R. Bonami (eds.).Atlas of invertebrate viruses. CRC press, Inc., BocaRaton, FL.

Anthony, D. W., M. D. Lotzkar and S. W. Avery.

1978a. Fecundity and longevity ofAnopheles albi'manus exporcd at each larval instar to spores ofNosema algerae. Mosq. News 38:l l6-121.

Anthony, D. W., K. E. Savage, E. I. Hazard, S. W-Avery, M. D. Boston and S. W. Oldacre. 1978b.Field tests with Nosema algeraeYarva and Undeen(Microsporida, Nosematidae) against Anopheles al-bimanus Wiedemann in Panama. Misc. Publ. En-tomol. Soc. Am. ll:.17-27.

Baumann, P., M. A. Clark, L. Baumann and A. H.Broadwell. 199 l. B acil lus sp haericus as a mosquitopathogen: properties ofthe organism and its toxins.Microbiol. Rev. 55:425-436.

Brooks, W. M. 1988. Entomogenous protozoa, pp.l-149. In: C. M. Ignoffo and N. B. Mandava (eds.).Handbook of natural pesticides, Volume V. Micro-bial insecticides, Part A. CRC Press, Inc., Boca Ra-ton. FL.

Chapman, H. C. 1985. Ecology and use of Coelo-momyces in biological control: a review, pp. 361-368. In: J. N. Couch and C. E. Bland (eds.). Thegenus Coelomomyces. Academic Press, New Yorkand London.

Collins, F. H. and N. J. Besansky. 1994. Vector bi-ology and the control of malaria in Africa. Science264:.1874-187 5

Couch, J. N. and C. E. Bland. 1985. Taxonomy, pp.82-297. In: The genus Coelomomyces. J. N. Couchand C. E. Bland (eds.). Academic Press, New Yorkand London.

Couch, J. N. and C. J. Umphlett. 1963. Coelomo-myces infections, pp. 149-188. In: E. A. Steinhaus(ed.). Insect pathology: an advanced treatise. Aca-demic Press, New York.

de Barjac, H. and D. J. Sutherland (editors). 1990.Bacterial control of mosquitoes and blackflies, Bio-chemistry, genetics and applications of Bacillus thu-ringiensis israelensis and, Bacillus sphaericus. Rl:ut-gers Univ. Press, New Brunswick, NJ.

Federici, B. A. 1974. Virus pathogens of mosquitoesand their potential use in mosquito control, pp. 93-135. In: A. Augin (ed.). Mosquito control. Univ. ofQuebec Press, Montreal.

Federici, B. A. 1981. Mosquito control by the fungiCulicinomyces, Lagenidium, and Coelomomyces, pp.555-572. In: H. D. Burgess (ed.). Microbial controlof pests and plant diseases 197f1980. AcademicPress, London.

Federici, B. A., P. Luthy and J. E. Ibarra. 1990. Theparasporal body of BTI: structure, protein compo-sition, and toxicity, pp. 1644. In:H. de Barjac ancD. Sutherland (eds.). Bacterial control of mosquitoesand blackflies; biochemistry, genetics, and applica-tions of Bacillrc thuringiensis and Bacillw sphaeri-cxc. Rutgers Univ. Press, New Brunswick, NJ.

Goldberg, L. J. and J. Margalit. 1977. A bacterialspore demonstrating rapid larvicidal activity againstAnopheles sergentii, Uranotacnia unguiculata, Culexunivitattus, Aedes aegypti, and Culex pipiens. Mosq.News 37:355-358.

Guillet, P. D,, C. Kurrak, B. Philippon and R. Meyer.1990. Use of Bacillus thuingiensis israelensis foronchocerciasis control in West Africa, pp. 187-201 .In: H. de Barjac and D. Sutherland (eds.). Bacterialcontrol of mosquitoes and blackflies; biochemistry,

268 Jouruar or nrE AvrnrclN Mosquro CoNrnol Assocreuoll V o l . l l . N o . 2

genetics, and applications of Bacillus thuringiensisand Bacillw sphaericus. Rutgers Univ. press, NewBrunswick. NJ.

Hall, R. A. and B. Papierok. 1982. Fungi asbiologicalcontrol agents ofarthropods ofagricultural and med-ical importance. Parasitology 84:205-240.

Henry, J. E. 1981. Natural and applied control ofinsects by protozoa. Annu. Rev. Entomol. 26:49-73.

Hougard, J.-M. 1990. Formulations and persistenceof Bacillw thuringiensis in Culex quinquefasciatuslarval sites in tropical Africa, pp. 295-306. In: H.de Barjac and D. Sutherland (eds.). Bacterial controlof mosquitoes and blackflies; biochemistry, genetics,and applications of Bacillus thuringiensis and Ba-cillus sphaericw. Rutgers [Jniv. Press, New Bruns-wick, NJ.

Jaronski, S., R. C. Axtell, S. M. Fagan and A. J. Dom-nas. 1 983. In vitro production of zoospores by themosquito pathogen Lagenidium giganteum on solidmedia. J. Invertebr. Pathol. 41:305-309.

Kerwin, J. L. and R. K. Washino. 1987. Ground andaerial application of asexual stage of Lagenidiumgiganteum for control ofmosquitoes associated withrice culture in the Central Valley of California. J.Am. Mosq. Control Assoc. 3:59-64.

Lacey, L. A. and A. H. Undeen. 1986. Microbialcontrol of blackflies and mosquitoes. Annu. Rev.Entomol. 3l:265-296.

Levy, R.and T. W. Miller, Jr. 1977. Experimentalrelease of Rorr anomer mi s cul icivor ax (Mermithidae:Nematoda) to control mosquitoes breeding in south-west Florida. Mosq. News 37:483486.

McCoy, C. W., R. A. Samson and D. G. Boucias. 1988.Entomogenous fungr, pp. l5l-236. 1n: C. M. Ignoffoand N. B. Mandava (eds.). Handbook of natural pes-ticides, Volume V. Microbial insecticides, Part A.CRC Press, Inc., Boca Raton, FL.

Merriam, T. L. and R. C. Axtell. 1982. Evaluationof the entomogenous fungi Culicinomyces clavos'porus and Lagenidium giganteum for control of the

salt marsh mosquito Aedes taeniorhynchus. Mosq.News 42:592[-602.

Mulla, M. S. 1990. Activity, field efrcacy, and use ofB acillus t huringiensis israelen^ls against mosquitoes,pp. l3zt-160. In: H. de Barjac and D. Sutherland(eds.). Bacterial control ofmosquitoes and blackflies;biochemistry, genetics, and applications of Bacillusthuringiensis and, Bacillus sphaericus. Rutgers Univ.Press, New Brunswick, NJ.

Petersen, J. J. 1982. Current status ofnematodes forthe biological control ofinsects. Parasitology 84:177-204.

Porter, A. G., E. W. Davidson and J.-W. Liu. 1993.Mosquitocidal toxins of bacilli and their genetic ma-nipulation for effective biological control of rnos-quitoes. Microbiol. Rev. 57:838-86 l.

Soares, G. G., Jr. and D. E. Pinnock. 1984. Effect oftemperature on germination, growth and infectivityof the mosquito pathogen Tolypocladium cylindros-porum (Dettetomycotina: Hyphornycetes). J. Inver-tebr. Pathol. 43:242-246.

Sweeney, A. W., E. I. Hazard and M. F. Graham. I 98 5.Intermediate host for an Amblyospora sp. (Micros-pora) infecting the mosquito, Culex annulirostris. J.Invertebr. Pathol. 46:98-1 02.

Sweeney, A. E., R. Cooper, B. E. Medcraft, R. C. Rus-sel, M. O'Donnell and C. Panter. 1983. Field testsof the mosquito fungus Culicinomyces clavisporusagainst the Australian encephalitis vector Culex an-nulirostris. Mosq. News 43:.29o-297.

Whisler, H. C. 1985. Life history of species of Coe-lomomyces, pp. 9-22. In; J. N. Couch and C. E.Bland (eds.). The genus Coelomomyces. AcademicPress, New York and London.

Yap, H.-H. 1990. Field tials of Bacillus sphaericusfor mosquito control, pp.3O7-320.1n.'H. de Badacand D. Sutherland (eds.). Bacterial control of mos-quitoes and blackflies; biochemistry, genetics, andapplications of Bacillus thuringiensis and Bacillussphacrictu. Rutgers Univ. Press, New Brunswick, NJ.