Large-scale applications of non-persistent but broad-spectrum chemical insecticides in Africa during the 1980s for control of acridoid pests, particularly the desert locust (Schistocerca gregaria Forskål) and Sahelian pest grasshoppers, raised concern about environmental damage and human safety. Similar concerns have been expressed in Australia, the United States, and Canada and have led to a search for alternative strategies. To lessen dependence on chemicals, an integrated pest management (IPM) approach for grasshopper control has been encouraged in the United States with emphasis on biological control as an important component and this is also desirable elsewhere, but additional biocontrol components are needed. Current strategies for most pest acridoids rely on short-term destruction of outbreak populations. Nymphs are the preferred target wherever possible and inundative augmentation of entomopathogenic deuteromycete fungi formulated as biopesticides could replace chemical spraying in some cases, especially where the major threat is to crops remote from the pest breeding areas. Entomopathogens are slower acting than chemicals and thus best suited for use where the pest is not immediately threatening to crops. Schistocerca gregaria and Oedaleus senegalensis Krauss pose particularly difficult problems because of the very large area and inaccessibility of their potential breeding grounds, their very sudden upsurges, and their great mobility as adult swarms. Fast-acting chemicals are likely to be needed when rapid intervention is required to control these pests, but an IPM strategy could incorporate biopesticide application in the early stages of upsurges and also be used for swarm control in some cases. However, improved prediction and monitoring are needed to facilitate the use of biopesticides and other IPM techniques against these pests.

A total of 181 isolates of Metarhizium anisopliae (Metschnikoff) Sorokin, M. flavoviride Gams and Rozsypal, Beauveria bassiana (Balsamo) Vuillemin, and Sorosporella sp. was found in a survey of Orthoptera in West Africa, Madagascar, Oman, and Pakistan between 1990 and 1993. Prior to this survey, there were only 28 isolates of hyphomycete fungi from Orthoptera held in international culture collections. Seventeen of the recently acquired Metarhizium isolates have been determined to be highly virulent during screening tests as part of a research programme for the development of a microbial insecticide against locusts and grasshoppers in Africa. Ninety-five isolates came from Benin which was the country where survey activities were most concentrated, and 63 of these isolates were found in Malanville, northern Benin, between 1991 and 1992 during an epizootic of M. flavoviride. Recordings from Oman and Pakistan represent the first specimens from these countries to be deposited in international culture collections. No deductions can be made on the best method for survey; both incubation of live grasshoppers and field searches for cadavers yielded results. Soil baiting with Orthoptera was used with some success. Limited soil screening using selective agar media was not found to be particularly useful.

The success of microbial control programmes often depends on an adequate mass-production method for the agent used. The pathogens with potential for use against grasshoppers and locusts vary widely in the ease with which they can be mass-produced. Obligate pathogens such as grasshopper entomopoxviruses, mermithid nematodes and Nosema locustae Canning are currently restricted to culture in living systems. Liquid fermentation is usually employed for the production of non-obligate pathogens such as bacteria, some fungi and nematodes, but in some cases the propagules produced in liquid culture are not amenable to formulation and application. Conidia of hyphomycete fungi can be produced on the surface of liquid media, on solid substrates or in diphasic systems. Production on solid substrates has been adopted for production of steinernematid and heterorhabditid nematodes and some fungi. Diphasic liquid-solid fermentation combines the benefits of both systems and is used mostly for mass-production of hyphomycete fungi such as Metarhizium flavoviride Gams & Rozsypal and Beauveria bassiana (Balsamo) Vuillemin. Increased commercial interest in biological control is likely to accelerate the development of improved and more economical methods for the mass-production of microbial control agents.

Successful development of a biological pesticide requires attention not only to the biological agent, but also to formulation, application, and the biology of the pest–pathogen interaction in the field. Emphasis in our review is given to fungi, Metarhizium spp. and Beauveria bassiana (Balsamo) Vuillemin, as the most suitable agents, and oil-based ULV formulations or baits as the most promising application techniques for use with locusts and grasshoppers. The efficacy of the pathogen isolate must be maximized; selection is aimed at those that are suitably virulent, specific, and well adapted to the relevant environmental conditions. Opportunities exist for manipulation of the characteristics of the isolate by genetic means and by developments in culturing techniques. Formulation requirements are stability during storage and the ability to carry the active ingredient successfully to the target insect at application. Likely storage methods for fungi would be as dry conidia, perhaps with clay diluents, or in oils; the characteristics of both are briefly discussed. At application, efficacy of dose transfer and protection of the biological agent against environmental constraints such as UV radiation are needed. Baits have advantages in terms of dose transfer but logistical problems associated with the bulkiness of the carrier remain. Technological advances, including those that offer the prospect of carrier production in situ from dense precursors, and better knowledge of feeding behaviour have improved the prospects for baits. Multi-disciplinary research reducing dependency on the biological agent and exploiting formulation chemistry and application technology is required in developing biological pesticides.

The use of chemical insecticides, especially as ultra low volume (ULV) formulations, against locusts and grasshoppers will continue for the foreseeable future; therefore application techniques for microbial agents should be as compatible as possible with existing practice. Low volume and ULV spraying of deuteromycete conidia in oil-based formulations have produced very promising acridid control results in the field, although baiting, dusting, and hydraulic application techniques have also been tested for a wide range of pathogens.The key problems for further research and development appear to be the logistics and supply of consistently reliable formulations for application on a large scale, and the determination of mechanisms for effective dose transfer in the field. The application of suspended particulate matter can present special problems with rotary and other atomizers.

Microbial control agents offer a method of pest control using organisms that are a natural component of the environment and are usually much more selective than chemical pesticides. Furthermore, they can usually be integrated with other methods of control, and may provide prolonged control by establishment within the host population. However, microbial control agents also possess properties that can pose human and environmental risks depending on the nature of the pathogen and its pattern of use. We present an overview of issues concerning the safety and registration of microbial control agents with emphasis on pathogens of locusts and grasshoppers. The potential safety issues and other consequences of concern from the deployment of microorganisms for pest control are: (1) pathogenicity to non-target organisms, (2) toxigenicity to non-target organisms, (3) competitive displacement of microorganisms, and (4) allergenicity. Inundative control methods pose unique risks because the pathogens must be produced in large quantities, stored, transported, and applied, usually in concentrations much higher than would normally ever occur naturally. The overriding concern in introducing an exotic agent is the risk to non-target beneficial organisms, because once the agent becomes established, it will in most situations be impossible to eradicate. However, if indigenous organisms are used, there is relatively little risk of irreversible, long-term detrimental effects. A synopsis of safety testing results of some of the more promising microbial control agents for grasshoppers and locusts and an evaluation of their potential hazards are presented. Safety to vertebrates is evaluated by a tiered series of laboratory test requirements. Assessments on hazards to non-target invertebrates are based principally on results of laboratory bioassays. Safety tests should be chosen with regard to the biological characteristics of the agent and should not impose standards that are more stringent than those imposed on other forms of pest control. Regulatory oversight should assure the integrity of the environment and safety of the public, while at the same time not unduly hampering the development, registration, and use of more sustainable pest control methods.

The net present value (NPV) approach to capital budgeting is used to determine the relative economic feasibility of two production models capable of manufacturing a fungi-based biopesticide in Madagascar. Sales revenues are projected at $10–12 per hectare for 20 000–80 000 ha annually, with recurrent costs estimated in Madagascar and investment costs from IITA (Cotonou, Benin) and Mycotech Corporation (Butte, Montana). These cash flows are discounted by an appropriate interest rate and risk factor, with positive results for both the labour-intensive model and the capital-intensive model under several scenarios. Cost advantages for the two models depend on both technology and scale. The labour-intensive model achieves a higher NPV in a market of 20 000 ha per annum as compared with the capital-intensive model. The capital-intensive model achieves a higher NPV in a market of 80 000 ha (including exports to southern Africa). Both models benefit from scale economies, although this benefit is relatively greater for the capital-intensive model. Consumers of mycopesticides in Madagascar could realize nearly 20% savings under a higher output scenario with a capital-intensive technology, than under a lower output scenario with a labour-intensive technology. Large-scale producers, however, would require nearly four times as much investment capital, and could find it difficult to produce for export from Madagascar. In the absence of a large-scale producer, small-scale production would be appropriate and feasible based on lower investment costs. Malagasy production is also protected from foreign competition because of current phytosanitary regulations.

Entomopoxviruses (EPVs) are insect poxviruses that are often found infecting grasshoppers and locusts. Nearly 15 grasshopper and locust EPVs have been reported in the literature. This review describes our current knowledge of the biology of grasshopper and locust EPVs including virus ultrastructure, host range, production in cell culture, pathology, process of infection, epizootiology, and field evaluations of the viruses to assess their potential as biological control agents. The most extensively studied has been the Melanoplus sanguinipes EPV (MsEPV). Trypsin-like protease activity has been identified in association with MsEPV occlusion bodies but its importance in the infection process is not known. Mortality from MsEPV has been found to occur in two distinct time frames over 6 weeks or longer. MsEPV is also the only grasshopper EPV that has been grown in vitro and been shown to produce virus that is both infectious and virulent to M. sanguinipes. Horizontal transmission of grasshopper EPVs is apparently by consumption of infected cadavers. Field evaluations of MsEPV at an application rate of 1 × 1010 occlusion bodies per hectare resulted in a 23% prevalence after 13 days despite a considerable amount of dispersal of grasshoppers between plots. Epizootiological studies of EPVs will continue to be an area requiring additional research. Virus production and a limited host range are the two most critical issues affecting the development of EPVs as microbial control agents.

Entomopoxviruses (EPVs) are large DNA viruses with structural similarities to vertebrate poxviruses. EPV virions are occluded in large (3–15 μm in diameter) proteinaceous occlusion bodies (OBs). To date, EPVs are reported from 15 species of grasshoppers and locusts. The current information on the biochemical characterization of these EPVs is summarized in our review. The DNA genomes of grasshopper and locust EPVs analysed to date have a G+C ratio of approximately 18.5% and genome size estimates generated by various methods range from 180 to 194 kilobase pairs (kbp). Restriction endonuclease enzyme analysis of a number of grasshopper and locust EPV DNAs shows the virus isolates to be distinct and the technique will be useful in identifying virus isolates. The structural proteins of certain grasshopper EPVs have also been analysed. Forty to 50 polypeptides ranging in molecular weight from 12 to 200 kilodaltons (kDa) have been detected by SDS-PAGE analysis of virions released from OBs and the polypeptide profiles are distinct for many of the virus isolates. The proteinaceous matrix of the OB of EPVs contains one predominant protein referred to as spheroidin. The spheroidin protein from most grasshopper EPVs is approximately the same molecular weight, 107 kDa, when analysed by SDS-PAGE. As with other groups of occluded insect viruses, grasshopper EPVs have a protease activity associated with OBs derived from infected insects. The possible role of this protease activity in the infection cycle is discussed. Finally, the role of various molecular techniques for the detection and identification of EPV infections in laboratory and field populations of grasshoppers and locusts is discussed. The development of EPV-specific monoclonal antibodies and DNA hybridization probes for the detection of virus infections is reviewed. As well, the possible use of polymerase chain reaction and randomly amplified polymorphic DNA fingerprinting techniques for the detection and identification of EPV infections is discussed.

Bacteria have been implicated in disease epizootics observed in field populations and laboratory-reared locusts and grasshoppers. Two species [Serratia marcescens Bizio and Pseudomonas aeruginosa (Schroeter) Migula] consistently infect locusts when ingested with food and can spread in laboratory populations. However, research on developing these organisms for microbial control of locusts and grasshoppers begun in the 1950s has not been continued. In recent years strains of Bacillus thuringiensis Berliner have been studied for activity against locusts and grasshoppers. Results of additional trials by the authors are reported. Among 393 B. thuringiensis isolates and 93 preparations of other sporeforming bacteria fed to nymphs of Locusta migratoria (L.) and/or Schistocerca gregaria Forsk., none has shown any pathogenicity to the insects. The recent discovery of novel B. thuringiensis strains active against various diverse pests and the many properties of a sporeforming bacterium that satisfy the requirements for a microbial control agent, and the development of Serratia entomophila as a promising agent for control of grass grubs, provide incentive to continue the search for an orthopteran-active sporeforming bacterium and to re-investigate the potential of non-sporeforming bacterial pathogens as microbial control agents of grasshoppers and locusts.

The relationship of nematodes and nematomorphs with grasshoppers and locusts is reviewed, emphasizing the actual or potential role of these parasitoids in microbial management. There are records of mermithids parasitizing grasshoppers worldwide, and they are considered important biological control agents in some grassland ecosystems of Europe, North and South America, Papua New Guinea, New Zealand, and Australia. Nematomorphs, although widely distributed, are uncommon parasitoids of grasshoppers, and their dependence on free-standing water for host infection and apparent host specificity are considered drawbacks to their use in biological control programmes. The ascaridids, spirarids, and acanthocephalans are parasites of birds and mammals, and may use grasshoppers as intermediate hosts. They have been shown to debilitate grasshopper hosts in laboratory studies, but their primary role as parasites of vertebrates precludes any consideration as biological control agents. Rhabditids do not naturally parasitize grasshoppers, but recent advances in mass-culturing techniques have given them a potential role as bioinsecticides for the control of grasshoppers. Quantitative data on the effects of nematodes and nematomorphs on agricultural pests, including grasshoppers and locusts, are generally lacking. However, there is evidence that some, particularly mermithids, are important in the population dynamics of grasshoppers and locusts. Keys to the identification of the various 'worms' found in grasshoppers and locusts are provided, including keys to the species of mermithids.

Several species of entomopathogenic deuteromycetous fungi can produce epizootics in populations of grasshoppers and locusts. Consequently there is considerable interest in development of these fungi as biocontrol agents. To this end we need information about the genetic and molecular basis of deuteromycete pathogenesis in acridids to develop a rational plan for strain improvement. Herein we present an overview of the infection processes of deuteromycetous fungi in acridids. These fungi penetrate through the cuticle which is composed primarily of proteins. Hydrophobic interactions, appressoria formation, and mucus production by the fungus are involved in fungal adhesion to the acridid cuticle. Extracellular proteases produced by Beauveria bassiana (Balsamo) Vuillemin and Metarhizium anisopliae (Metchnikoff) Sorokin solubilize cuticle proteins, which assists penetration and provides nutrients for further growth. Fungal infection through the locust gut is rare because indigenous gut microflora produce antifungal metabolites. Little is known of the events providing host specificity or those that lead to insect death once the cuticle is breached by the fungus; however, mechanical damage, nutrient deprivation, and toxic metabolites may be involved.

Recognition of the potential of Beauveria bassiana (Balsamo) Vuillemin as a control agent of grasshoppers and locusts occurred as early as 1936, in South Africa. Field testing of B. bassiana as an inundative control agent of grasshoppers and locusts has been facilitated by development of a solid substrate method for mass-production of the fungus and has resulted in the registration of a strain against grasshoppers in the United States. In some, but not all field trials, application has resulted in substantial reductions in grasshopper populations. Numerous environmental constraints, including temperature and ultraviolet (UV) radiation, may limit field efficacy of the fungus. Laboratory studies suggest that low humidity does not limit the ability of the fungus to initiate disease. Sunlight is the major cause of mortality of conidia on leaf surfaces. The incorporation of UVB protectants in formulations can increase conidial survival; however, these have not yet been evaluated for their effects on field efficacy of B. bassiana against insects. Thermoregulation by grasshoppers has been implicated in resistance to mycosis. Results of laboratory studies indicate that grasshoppers infected with B. bassiana preferentially seek temperatures between 40 and 42 °C and these temperatures are inhibitory to disease development. In field-cage trials, a higher prevalence and more rapid development of disease were observed in grasshoppers placed in shaded cages than in grasshoppers placed in cages exposed to full sunlight. In laboratory experiments simulating grasshopper thermoregulation during daylight periods, application of both Metarhizium flavoviride Gams and Rozsypal and B. bassiana simultaneously resulted in a final prevalence of disease that was greater than M. flavoviride alone in the hot temperature environment, and equal to B. bassiana alone in the cool temperature environment. Incorporation of sublethal levels of Dimilin with conidia of B. bassiana increased efficacy of the fungus against grasshoppers in laboratory and field trials. Once environmental constraints are better quantified, it may be possible to overcome them through improved formulation, strain selection, genetic or phenotypic manipulation, and inoculum targeting. Ultimately, success of B. bassiana as a microbial control agent will depend on our ability to overcome environmental and other constraints and/or to predict its efficacy under various environmental conditions.

Two isolates of the fungus Beauveria bassiana (Balsamo) Vuillemin, GHA and BF, were evaluated in Cape Verde in 1991 and 1992 for infectivity to the Senegalese grasshopper, Oedaleus senegalensis (Krauss), and the migratory locust, Locusta migratoria migratorioides (Reiche and Fairmaire). Evaluations included laboratory bioassays and small-scale field trials. Laboratory bioassays evaluated five different formulations. Four of the formulations tested showed strong dose–response patterns and significantly higher mortality than the untreated control or carriers minus spores. All four formulations achieved high mortality levels when applied at economically feasible dose rates. The GHA and BF isolates, formulated in an oil carrier with an emulsifier, were equally infectious to migratory locust nymphs. Six different formulations of GHA were evaluated in field trials. Field trials evaluated both direct effects (treatment of field plots infested with O. senegalensis) and indirect effects (treatment of plots without grasshoppers, after which grasshoppers were introduced). In both cases, all six formulations showed good biocontrol potential. Grasshoppers exposed to treated plots up to 72 h after application exhibited comparatively high mortality levels, indicating that large numbers of spores remained viable in the field for at least 3 days. This was confirmed by analysis of the viability of conidia from vegetation samples obtained in the field following treatment. In open-plot, small-scale field trials, two different formulations (oil and clay-based) of GHA resulted in high rates of infection and approximately 45% reductions in grasshopper densities in the treated plots 7 days after application, even though density-reduction results were "diluted" by grasshopper migration into and out of the test plots. Results of the Cape Verde evaluations demonstrate that biopesticides developed from B. bassiana represent a promising alternative to chemical pesticides for grasshopper and locust control.

The influence of three formulations, water, oil, and a 5% oil emulsion, and two crops, alfalfa and crested wheatgrass, on the deposition and subsequent persistence of Beauveria bassiana (Balsamo) Vuillemin conidia in soil was investigated. The alfalfa canopy was considerably denser than that of wheatgrass. Leaf area indices for alfalfa ranged from 1.8 to greater than 2, those for wheatgrass ranged from 0.24 to 0.55. Initial populations of conidia averaged 1.2 × 103 to 2.6 × 104 colony-forming units (cfu) per gram of dry weight of soil under alfalfa, and 5.5 × 103 to 3.4 × 104 cfu per gram of soil under wheatgrass. There was no consistent influence of formulation or application method (high or ultra low volume) on penetration of conidia through the canopy of either crop. However, conidial populations under wheatgrass were larger than those under alfalfa in two of three trials. After 225–272 days (over winter), substantial populations (87 to 4.3 × 104 cfu/g) were recovered from soil. Although conidial densities decreased over time, reductions in population size over this period were generally less than one order of magnitude; neither crop nor formulation consistently influenced conidial persistence. In most instances, a rapid decrease in conidial populations was observed within approximately 20 days but thereafter, the rates of population decline abated. The initial decrease in conidial numbers did not appear to be related to precipitation. This study demonstrates that substantial numbers of B. bassiana conidia infiltrate crop canopies, are deposited on the soil surface, and subsequently persist in a clay–loam soil. The aerial application of B. bassiana conidia to vegetated roadsides may prove useful for the management of ovipositing grasshoppers and emerging nymphs.

Four research programmes are investigating the entomopathogenic fungal genera Metarhizium and Beauveria for locust and grasshopper control in Africa. In the LUBILOSA programme, surveys for pathogen isolates revealed a morphologically distinctive Metarhizium flavoviride Gams and Rozsypal attacking acridoids in West Africa, Madagascar, and elsewhere. Metarhizium anisopliae (Metschnikoff) Sorokin isolates with virulence to acridoids were also obtained, including several from non-orthopteran hosts. Natural epizootics of both genera are rare in acridoid populations, but do occur. A standardized screening method discriminated virulent from non-virulent isolates. The great majority of the most virulent isolates were from the acridoid group of M. flavoviride. A Niger isolate chosen for development from this group had low virulence to honey bees and parasitic Hymenoptera and was not infective to insects in several other orders. Field tests were carried out on formulations of oil mixtures, using ULV application rates of 1–2 L/ha and 2–5 × 1012 conidia per hectare. In preliminary tests, target insects were sprayed successfully in small field arenas and in large cages. Trials in 1993 on variegated grasshopper gave an approx. 90% reduction in field populations after 15 days. Trials on various acridids, predominantly Hieroglyphus daganensis Krauss, in dense grass in northern Benin showed slower mortality, although up to 70% population reduction was achieved. Trials using a vehicle-mounted ULV sprayer (the Ulva-Mast) in open grassland in Niger gave >90% mortality in samples of mixed acridids. In Mali, a Malian isolate of M. flavoviride was shown to be slightly more virulent than the standard Niger isolate; both isolates gave significant population reductions against nymphs of Oedaleus senegalensis Krauss and Kraussella amabile (Krauss) in 1-ha plots. Successful small-scale field trials have also been carried out using the standard M. flavoviride isolate in South Africa against brown locust and in Australia using an Australian isolate against wingless grasshopper. In Mauritania, a trial using the Niger isolate against desert locust nymph bands gave up to 90% mortality in caged samples by day 9 after spraying. The uncaged treated bands were completely destroyed by predators while untreated bands fledged. Oil-based ULV formulations of M. flavoviride are capable of causing high mortality in the field populations of all acridoids against which they have been field tested and show great promise for development as components of IPM strategies for these pests.

Only one isolate of Metarhizium flavoviride Gams and Roszypal group 3 has been isolated from a field-infected acridid in Australia. This is isolate FI985 (ARSEF 324) obtained from a spur-throated locust, Austracris guttulosa (Walker), near Rockhampton, Queensland, in 1979. In terms of conidial size and shape as well as phialide morphology, FI985 is intermediate between Metarhizium anisopliae (Metschnikoff) Sorokin and M. flavoviride. It has been compared with other group 3 isolates using RAPDs and sequence analysis of the ITS region and found to be very similar. However the analysis shows that these group 3 isolates are genetically closer to M. anisopliae than to M. flavoviride sensu stricto. Laboratory bioassays have shown that FI985 is virulent for five species of acridid pests in Australia. Comparative bioassays with other isolates of Metarhizium, including other group 3 isolates from Africa and Asia, have not yet revealed any isolate more virulent than FI985. This isolate is amenable to mass-production on rice and has been formulated in oil as a mycoinsecticide. The results from six field tests, mostly against wingless grasshopper, Phaulacridium vittatum (SjÖstedt), using doses of 2–7 × 1012 conidia per hectare and plot sizes up to 50 ha are summarized. These trials (with the exception of the first against the Australian plague locust) have given high levels of disease-related mortality in caged samples of the target collected within 3 days of spraying. In the four trials with wingless grasshopper, population reductions were detected 10–30 days after application; however these reductions were much less than suggested by cage samples as a result of movement of the target acridids. In contrast, positive control plots sprayed with fenitrothion gave a very high initial kill (>90% in 1 day) but were then more rapidly reinvaded. Consequently, 3–4 weeks after spraying the density in the plots treated with chemical insecticide and those treated with mycoinsecticide were similar. Further field trials are needed especially against the Australian plague locust and evaluating lower doses. The results obtained to date show that a mycoinsecticide based on FI985 is likely to be effective over a wide range of target acridids and weather conditions.

Trials on the use of Metarhizium flavoviride Gams and Roszypal conidia in oil-based formulation for the control of grasshoppers, particularly Hieroglyphus daganensis Krauss, in Malanville, north Benin, are described. Preliminary work examined sprayer types, application rate, and time of application. In a trial on 4-ha plots with three replicates, M. flavoviride mycoinsecticide application to H. daganensis nymphs resulted in field population reductions of 70% after 14 days. In samples incubated in cages, mortality was higher in the samples taken 3 or 7 days after application than in the sample taken immediately after application, indicating the possibility of residual pick-up compared with direct spray impact in this environment. Significant mortality was still being observed in samples collected 37 days after application; to investigate this further, a method for bioassaying the spore load in the field was developed and used to monitor the spore load in the field. The possibility that the results indicate the occurrence of secondary infections resulting from horizontal transmission of M. flavoviride is discussed.

A new isolate of Metarhizium flavoviride Gams and Rozsypal (Hyphomycetes) (CG 423) found in Northeast Brazil infecting Schistocerca pallens (Thunberg) was identified using arbitrarily primed PCR. Cluster analysis of DNA markers revealed a high level of homogeneity (>83% similarity) among the Brazilian (CG 423) and two other M. flavoviride isolates from Nigeria (CG 366 = IMI 330189) and Australia (CG 291). However, M. flavoviride isolates were very distinct when compared with two isolates of Metarhizium anisopliae (Metschnikoff) Sorokin (6.4% similarity). Bioassays showed that strain CG 423 is as virulent as other isolates of M. flavoviride (CG 291, CG 366), M. anisopliae (CG 087), and Beauveria bassiana (Balsamo) Vuillemin (CG 425) against the grasshopper Rhammatocerus schistocercoides (Rehn) (Orthoptera: Acrididae), an important pest in Central Brazil. However, the Brazilian isolate of M. flavoviride (CG 423) is more virulent than the Brazilian isolate of B. bassiana (CG 250). Because conidia used in bioassays were formulated in soybean oil containing 5% kerosene, the effect of the kerosene present in the oil formulation was tested. Kerosene (0–10%) did not affect the virulence (P > 0.3) of M. flavoviride against R. schistocercoides. The native isolate of M. flavoviride (CG 423) is now being developed as a mycoinsecticide against grasshoppers in Brazil.