Effect of two dosages of Metarhizium anisopliae var . acridum against Rhammatocerus schistocercoides Rehn

The fungus Metarhizium anisopliae var. acridum, strain CG 423, was tested under field conditions against the gregarious grasshopper Rhammatocerus schistocercoides (Rehn) (Orthoptera: Acrididae). Conidia formulated in a racemic mixture of soybean oil and kerosene were sprayed under field conditions using an ultralow-volume hand-held atomizer Ulva Plus adjusted to deliver 2.9 L/ha. Bands composed of 2 instar nymphs were treated with either 5.0x10 or 1.0x10 viable conidia/ha. The number of insects in each band was estimated at day one following spraying and by the end of the field trial (15 to 16 days posttreatment). Reductions in population size reached, in average, 65.8% and 80.4% for bands treated with the higher and lower dosage, respectively. For both dosages, total mortality rates of insects collected at two days post-application, and kept in cages for 14 days under lab conditions, showed no significant differences as compared to that obtained with insects collected immediately after spraying. Healthy insects were fed to native grasses sprayed on the field with 1.0x10 viable conidia/ha. Mortality levels of the nymphs fed on grasses collected two and four days post-application were not affected when compared to nymphs fed on grasses collected immediately following application.


Introduction
Although biopesticides represent approximately 1.3% of the phytosanitary products used in the world, those directed to insect control have 4.5% of the market share for insecticides (Menn & Hall, 1999).Products based on the bacterium Bacillus thuringiensis account for most of the sales, but those based on entomopathogenic fungi, specially the mitosporic group, have become very promising.In Brazil, production methods of entomopathogenic fungi were developed in the 70´s (Aquino et al., 1977), and are based on rice as the substrate.The yields are relatively low and, for fungi such as Metarhizium anisopliae var.acridum and M. anisopliae var.anisopliae, vary from 2.5 to 6.0x10 9 conidia per gram under laboratory conditions (Quintela, 1994;Magalhães & Frazão, 1996).As an additional disadvantage, conidia produced are not formulated, reducing the potential for longer shelflives and, therefore, turning production dependent on immediate demands.
In recent years, insect resistance to chemicals, high costs of research and development of new synthetic molecules, and technological breakthroughs on production, formulation and delivery of mycoinsecticides, have expanded the commercial success of these products (Wraight & Carruthers, 1999).For example, the private company Mycotech (Butte, Montana, USA) has developed a diphasic fermentation system that allows an average yield of 2.6x10 10 conidia of Beauveria bassiana per gram on a proprietary substrate (Johnson et al., 1992).In addition, formulations based on oils (Thomas et al., 1996;Lomer et al., 1997;Milner et al., 1997) and adjuvant emulsifiable oils (Alves, 1999;Bateman & Alves, 2000) have proved suitable in many circumstances.Despite the mentioned advances, the balance between economically feasible rates and satisfactory control levels is delicate and, in some situations, responsible for failure cases of biological control programs based on entomopathogenic fungi.Studies carried out at field conditions aiming the establishment of appropriate dosages are critical in biological control programs involving biopesticides.
Until recently, the gregarious grasshopper Rhammatocerus schistocercoides was considered a severe pest in the Chapada dos Parecis, Mato Grosso State, Brazil, attacking rice, corn and sugarcane (Lecoq & Pierozzi Junior, 1995a, 1995b;Miranda et al., 1996).This pest was chosen as a model system for a research project started in 1993 and aiming the microbial control of grasshoppers.Preliminary studies resulted in selection of the fungal pathogen M. anisopliae var.acridum, strain CG 423 (Moreira et al., 1996;Magalhães et al., 1997b) for deeper studies.A recent field trial performed at the permanent outbreak zone of R. schistocercoides resulted in outstanding mortality rate (88%), demonstrating the fungus to be well suited for control strategies of grasshopper pests (Magalhães et al., 2000b).However, it was sprayed at a very high dosage, equivalent to 2.0x10 13 conidia/ha.

Fungal production and formulation
Conidia of the strain CG 423 of M. anisopliae var.acridum were produced in parboiled rice according to the methodology described in Magalhães & Frazão (1996).Conidia were transported to Campos de Júlio county (Mato Grosso State) as a powder, and added to a racemic mixture of soybean oil plus kerosene prior to field application.

Selection of grasshopper bands
Search for grasshopper bands was carried out for three days before first fungal application.Different sites were visited and all bands mapped.The accurate position of each band was recorded on a Global Positioning System.Nine bands, considered as having adequate sizes (between 120 and 600 m 2 ), and located in areas far away from roads and, as much as possible, distant from other bands were selected.Information regarding the experimental area and procedures related to monitoring of grasshopper bands were previously reported (Magalhães et al., 2000b).

Bioinsecticide application
Red nozzle output and deposition tests using a handheld ULV spinning-disk atomizer, powered by six 1.5 V batteries, indicated the output to be 70 mL per minute, and the adequate spacing between swatches to be 4 m.Walking speed of 1 m/s and emission height of 1 m were adopted.Volume application rate was equivalent to 2.9 L/ha, and field dosages of 5.0x10 12 or 1.0x10 13 viable conidia/ha.For each dosage, 3 bands with most nymphs being on the 2 nd instar were treated.Bands treated with the low dosage (5.0x10 12 ) were designated as L1, L2 and L3, and those treated with the high dosage (1.0x10 13 ) were referred to as bands H1, H2 and H3.Spraying was limited to the surface area covered by each band, ranging from approximately 100 to 300 m 2 at the time of spraying.Untreated bands (U1, U2 and U3) were considered as controls.Table 1 indicates the prevailing conditions by the time spraying was performed.Prior to application, conidial viability was determined according to methodology described by Magalhães et al. (1997a), and re-assessed following spraying.Bands were daily followed in their natural environment, from 15 (low dosage bands) to 16 days post-spraying (high dosage bands).

Measuring size and density of bands
Surface area and total number of nymphs per band were estimated at the beginning (day one) and by the end of the field trial.Surface area was determined by delimiting the band and dividing it in four different estimated density zones at day 1 post-spraying.The densities were 2,512, 696, 192, and 39 nymphs per m 2 , considered as extremely high, high, medium and low, respectively.Average numbers were obtained by throwing a metal barrel measuring 0.26 m 2 on specific density zones within bands and counting the number of captured insects.The total number of nymphs per band was determined by multiplying the surface area of each density zone by the correspondent number of nymphs per m 2 , and adding the values up.Since the average number of nymphs per density zones changed according to nymphal instar, number of individuals in a given band was overestimated by the end of the experiment, when band sizes had increased due to molting of nymphs to 3 rd instar.Nymphal mortality due to natural factors in control bands was expected to be low, equivalent to a theoretical value of 0.86% a day (Miranda et al., 1996), which turns out to correspond to 13.8% after 16 days.This value was confirmed in a field trial carried out in 1998 (Magalhães et al., 2000b), when a mortality rate of 13.1% was observed after 14 days.Numbers calculated at the end of the field trial for treated bands were adjusted considering the expected natural mortality (0.86% per day) and the increasing size constant (ISC) derived from observations of untreated bands.The constant (ISC = 2.19) was obtained by dividing the average number of insects in untreated bands estimated at the final and beginning of the trial.Distances daily traveled by bands were measured as well.

Caged insects
For each conidial dosage, nymphs were collected at zero, two and four days post-spraying of the fungus.For untreated bands, insects were collected only at the beginning of the experiment.Groups of 30 insects were kept in cages and fed to native grass Andropogon selloanus Hackel, which was replaced every other day.Insects were daily observed and dead insects placed on a moist chamber in order to confirm if death was caused by M. anisopliae var.acridum.Insects showing the typical red coloration or external mycelial growth on the cadaver were considered as killed by the pathogen.
In a second experiment with caged insects, the persistence of the fungus M. anisopliae var.acridum when sprayed on vegetation at a dosage equivalent to 1.0x10 13 conidia/ha was studied.Grass leaves collected at a sprayed site, immediately after, two and four days following spraying, were offered to healthy nymphs kept in cages.Total percentage of deaths and confirmed mortality rates were then determined.Each treatment comprised three replicates, and insects were kept under laboratory conditions for 14 days.In the last experiment, nymphs were collected 11 days post-spraying within spots of band H1/H2 presenting extremely high, high, medium and low nymphal densities.Insects were kept under laboratory conditions for seven days, then the cumulative total and confirmed mortality levels were recorded.Only one replicate was performed per class of nymphal density.
Table 1.Parameters prevailing during spraying of fungus Metarhizium anisopliae var.acridum against the hopper Rhammathocerus schistocercoides under field conditions.
(2) Numbers before and after slash refer to prevailing conditions at the beginning and at the end of the spraying period, respectively.

Analyses
Although there were three treated bands for each application rate at day one, the number of bands was reduced to two due to merging of bands along the period.Therefore, statistical analyses were not performed for the field trial.Regarding laboratory assays, percentage data were submitted to one way analysis of variance, and average comparisons were done with t test or Tukey at a 5% significance level.The software SigmaStat (Kuo et al., 1992) was employed for the statistical procedures.

Results and Discussion
One day prior to application, viability of pure conidia of Metarhizium anisopliae var.acridum was 89%.Following spraying, viability of conidia formulated in the racemic mixture of soybean oil and kerosene was re-assessed and showed no reduction.
No significant differences in total mortality were observed immediately following application, two or four days post-spraying (t test; P≥0.145) when the two dosages were compared under laboratory conditions (Table 2).A different trend was observed for confirmed mortality, since at immediately following application and two days post-application the averages for treatment with 5.0x10 12 conidia/ha were significantly higher (P≤0.038)than four days post-application.
For the low dosage, total mortality average of nymphs collected at day four post-spraying was significantly reduced (Tukey; P = 0.004).For the high dosage, no differences were observed among averages (P = 0.330).Regarding confirmed mortality, for both dosages the differences among averages were not significant (P≥0.080).According to Vicentini & Magalhães (1996), events related to attachment, germination and penetration of strain CG 423 into the host take place in 18 hours.The results demonstrate that the invasive capability of conidia exposed to environmental constraints for at least two days is not being severely affected.
No significant differences in either total mortality or confirmed mortality under laboratory conditions were found when nymphs were fed on grasses collected immediately following application, two and four days post-spraying (P = 0.167 and 0.056 for total and confirmed mortality, respectively), although there is a trend of grasses collected at day four to induce lower mortality levels (Table 3).During the trial period and, specially, until two days post-spraying, UV radiation and other detrimental factors did not play a major role in conidia inactivation.The oil formulation may have some contribution to this persistence, as previously reported in other studies (Moore et al., 1993;Alves et al., 1998).Besides, climatic condition at Campos de Júlio county during the field trial, characterized by mild temperatures, high relative humid-Table 2. Percentage of mortality of the hopper Rammatocerus schistocercoides nymphs collected at various times following field application of fungus Metarhizium anisopliae var.acridum (5.0x10 12 and 1.0x10 13 conidia/ha) and kept under laboratory conditions for 14 days.
(1) Insects from replicates 1, 2 and 3 of treatments 5.0x10 12 and 1.0x10 13 were collected from bands L1, L2, L3 and H1, H2 and H3, respectively. (2)In a given column, average numbers±standard error followed by the same letter are not significantly different according to t test (α = 0.05); average total and confirmed mortality rates for the control were 21.1±7.76%and 0%, respectively. (1)Grass leaves were collected immediately after application of the fungus and kept overnight at room temperature for 12 hours before offering to insects. (2)Data is missing (insects escaped from cage at day nine). ( 3 Within the same mortality parameter, average numbers±standard error followed by the same letter are not significantly different according to one way ANOVA.ity and intense cloudiness (Table 1), was an important factor that contributed to the satisfactory performance of the fungus.The persistence of M. anisopliae var.acridum on the foliage probably increases the level of fungal infection due to secondary pickup, but so far the role of this process was not addressed in our trials.
In the field, one band of the low dosage (L1) merged with an untreated band on December 2 nd and it was not considered thereafter.Two treated bands of the high dosage (H1 and H2) merged three days following spraying, and were considered altogether since then.The merging of bands is a natural phenomenon in areas presenting high density of bands, explained by their considerable mobility.In some instances, 2 nd instar nymphs were recorded moving as fast as 100 m a day.
Population reduction of R. schistocercoides nymphs at the end of the experiment ranged from 81.7 to 84.1% for the low dosage bands (L1 and L3) and from 61.6 to 79.6% for the high dosage bands (H1/H2 and H3).Interestingly, the light rain during application time of bands L1 and L2, and wet vegetation during spraying of bands L3 and H1, did not hamper the achievement of satisfactory control levels.In contrast to studies in which water-based formulations were used (Inglis et al., 1995), precipitation seems not to considerably reduce the attachment of oil-formulated conidia to hydrophobic surfaces, specially grasshopper cuticles and grass parts.An illustrative diagram of the structural changes in an untreated and treated bands is shown in Figure 1.
The values are close to the 88% reduction obtained in the 1998 field trial when a dosage equivalent to 2.0x10 13 conidia/ha was applied (Magalhães et al., 2000b).The results clearly suggest that economically feasible dosages (a quarter of that used in 1998) have a significant impact on grasshopper populations.Different studies have shown the possibility of achieving high control levels under field conditions applying dosages as low as 1.0x10 12 to 3.0x10 12 conidia/ha (Hooper et al., 1995;Price et al., 1997).Worldwide, the use of fungi for regulation of grasshopper populations is a successful biological control strategy and reviews on this subject can be found elsewhere (Goettel, 1992;Lomer, 1997;Magalhães et al., 2000a).Reports claiming successful control of other pests using field rates lower than 5.0x10 12 conidia/ha are rare, mainly in open field environments, moreover in semi-arid regions.
To date, it is not known if R. schistocercoides is able to thermoregulate (capability of controlling body temperature through active physiological or behavioral mechanisms), as observed for Melanoplus sanguinipes (Boorstein & Ewald, 1987), Taeniopoda eques (Whitman, 1988) and many other species (Chappell & Whitman, 1990).It is clear that during the rainy season, when the nymphal stage occurs and maximum temperatures are around 30ºC, any increase in body temperature is not being able to deter the fungal infection.In fact, during the peak of the epizootics, usually from 11 to 14 days post-spraying, dead insects were found in stem grasses in a basking position.
In the field trial the distances daily traveled by treated nymphs, when compared to untreated insects, were not severely affected.During the first eight days, average distances daily traveled in the control, and treatments with 5.0x10 12 and 1.0x10 13 conidia/ha were 43.0±4.77,38.5±2.90 and 53.3±16.75m, respectively (Figure 2).From eight days post-spraying to the end of the experiment, nymphs from the ).The authors also observed that treated insects were less mobile than healthy ones and usually succumbed to fungal infection before the 10 th day.There are some reasons that may explain the relatively long distances moved by treated nymphs under field conditions: conidial dose effectively delivered to each nymph is much lower than 5,000; and, nymphs were always collected in the front line of the bands, characterized for the high density of fast moving grasshoppers.The sampling procedure was biased toward fast moving nymphs, the ones that were less affected by the fungus.In order to solve this doubt, nymphs were collected 11 days post-spraying within spots of band H1/H2 with extremely high (front line), high, medium and low nymphal density (Table 4).Mortality rates observed in medium and low density spots were considerably higher, corroborating the hypothesis that sick nymphs, incapable of following healthy ones, tend to be found in low density zones (the ones located in the back of the band). (1)Densities zones extremely high, high, medium and low correspond to 2,512, 696, 192 and 39 nymphs per m 2 , respectively.
In fact, in the second half of the trial it was common to find groups of sluggish nymphs, not responsive to humans approximation.It was possible to observe a high number of spiders, ants and Calosoma sp.(Coleoptera: Carabidae) preying on R. schistocercoides nymphs.In the last days there were hundreds of reddish dead bodies on tracks of the treated bands, whereas this did not happen in the control treatment.
This study provides additional evidences for use of M. anisopliae var.acridum as mycoinsecticide for grasshopper control.Field trials to test other dosages and the impact of the fungus on non-target organisms are under way.

Conclusions
1. Nymphs of R. schistocercoides are susceptible to field dosages of M. anisopliae var.acridum as low as 5.0x10 12 conidia/ha.2. Mortality rates for nymphs collected up to two or four days post-spraying are not affected when the dosages 5.0x10 12 conidia/ha and 1.0x10 13 conidia/ha are applied.
3. High mortality rates for nymphs fed to grasses treated with M. anisopliae var.acridum, up to four days post-treatment, reinforce the argument that persistency of the fungus in the field is not severely affected by adverse environmental conditions at Campos de Júlio, MT.

Figure 1 .
Figure 1.Illustrative diagrams of untreated (band U1) and fungus-treated (L3) bands of Rhammatocerus schistocercoides at the beginning (day 1) and end (day 15) of the field trial.

Table 4 .
Percent mortality rates of the hopper Rhammatocerus schistocercoides nymphs collected in different density zones within a treated band at 11 days postspraying with fungus Metarhizium anisopliae var.acridum and caged for seven days.