SELECTION OF Beauveria bassiana ( BALS . ) VUILL . AND Metarhizium anisopliae ( METSCH . ) SOROK . FOR THE CONTROL OF THE MITE Mononychellus tanajoa ( BONDAR )

The green mite, Mononychellus tanajoa (Bondar) (Acari: Tetranychidae), is considered to be one of the key pests in cassava, Manihot esculenta Crants, leading to considerable field losses. In this study, ten Beauveria bassiana (Bals.) Vuill. and ten Metarhizium anisopliae (Metsch.) Sorok. isolates were evaluated with regard to their potential as biological control agents against adult M. tanajoa females. The total mortality percentage of M. tanajoa caused by B. bassiana ranged from 13.0 to 97.0%, with confirmed mortality rates extending from 9.0 to 91.0% and LT 50 varying from 4.2 to 17.0 days. The M. anisopliae isolates showed total mortality percentages ranging from 12.0 to 45.0% with confirmed mortality rates extending from 8.0 to 45.0%, and LT 50 varying from 8.6 to 19.8 days. Lethal Concentrations (LC 50 ) of 3.93 × 10 conidia mL and 7.44 × 10 conidia mL were determined for B. bassiana and M. anisopliae, respectively. B. bassiana isolate 645 was the most efficient, being an alternative for use in biological control programs against the cassava green mite.


INTRODUCTION
Brazil is one of the world's greatest cassava (Manihot esculenta Crantz) producers, with an annual yield of over 20 million tons, which places this crop among the main agricultural products explored in the country (FNP Consultoria e Comércio, 2002).Depending on the use of different technologies and management practices, cassava productivity can vary from 8 to 25 t/ha (Silva & Santos, 2000).The Brazilian Northeastern Region stands out as the greatest produc-ing region, with a 59% share of the country's cassavagrowing area and 46% of the country's yield, on average (Cavalcanti & Araújo, 2000).In different producing regions, the crop is infested by a large number of arthropods, some of which cause considerable damage, as, for example, Erinnyis ello ello L., Phenacoccus spp., and Mononychellus tanajoa (Bondar) (Farias, 1991;Bellotti et al., 1999).Among these, M. tanajoa constitutes one of the major problems for the crop in the Northeast Region (Moraes & Flechtmann, 1981).The cassava green mite M. tanajoa occurs preferably in the apical bud of the plant.Infested leaves become chlorotic, do not attain normal development, with a usually distorted growth, and may fall progressively from top to bottom.Attacked stalks lose their green color and become coarse and brownish, with a furrowed bark.The plant may completely exhaust its reserves and die (Flechtmann, 1989;Gallo et al., 2002).In the back country of the State of Pernambucano, infestations occur with greater intensity from July to November, and may cause root yield losses in the order of 51% (Veiga, 1985).
The use of entomopathogenic fungi is an important practice that should be incorporated into integrated management, with the objective of reducing the populations of pest insects and mites in economically important crops (Alves, 1998).Within this context, Odongo et al. (1998) mentioned the potential for controlling M. tanajoa using the entomopathogenic fungus Hirsutella thompsonii Fisher, as well as the use of Entomophthora spp., Beauveria bassiana (Bals.)Vuill., and Metarhizium anisopliae (Metsch.)Sorok.(Odindo, 1992).
In a paper involving different control methods against the mite Brevipalpus spp. in Mexico, Acevedo & Rosas (2000) concluded that control with H. thompsonii was superior to organophosphate chemical products.The fungus significantly reduced populations and the damage caused by the mites Brevipalpus phoenicis Geijskes and B. obovatus Donnadieu in citrus.Neozygites sp.epizootics are also frequent in M. tanajoa populations (Delalibera Jr. et al., 1992;Yaninek et al., 1996;Elliot et al., 2000).Several studies involving this fungus have been conducted; however, differently from M. anisopliae and B. bassiana, its challenging production on artificial medium makes it difficult to use it in green mite biological control programs (Oduor et al., 1996;Leite et al., 2000).
The fungi B. bassiana and M. anisopliae have been studied recently for the control of tetranychid mites (Alves et al., 2002;Oliveira et al., 2002).Notwithstanding, these entomopathogens have not yet been studied for the control of the cassava green mite.Hence, the present research had the objective of evaluating the effect of different B. bassiana and M. anisopliae isolates on M. tanajoa in the laboratory.

MATERIAL AND METHODS
The experiments were carried out in Recife -PE, Brazil.

Mite rearing
The rearing of Mononychellus tanajoa colonies in the laboratory was developed on cultivar Santo Estevão cassava plants, Manihot esculenta, planted in plastic pots containing soil and cattle manure at a 2:1 ratio.One 10cm plant cutting was placed in each pot; pots were maintained in the greenhouse.After 20 to 30 days, the plants were taken to the laboratory to be infested with mites.This process was repeated every seven days to maintain plant quality and to ensure a stock colony of mites.The pots containing infested plants were maintained in the laboratory at a temperature of 26 ± 1°C, 63 ± 5% relative humidity, and 12-hour photophase.

Obtaining and production isolates
The Metarhizium anisopliae and Beauveria bassiana isolates came from the fungus culture collection of the Insect Pathology Laboratory at the Plant Protection Division of UFRPE (Table 1).The isolates were plated into Petri dishes containing potato-dextrose-agar + streptomycine sulfate and incubated in a B.O.D. chamber at 26 ± 1°C and a 12-hour photophase for seven days.Next, they were reinvigorated on third-instar Diatraea saccharalis (F.) caterpillars in order to maintain the pathogens' virulence.The isolates were later stored in glass vials containing PDA culture medium and Nujol ® oil; in order to be used in the experiments, they were multiplied in plates containing complete culture medium (CM), consisting of yeast extract, glucose, minerals, agar, and water, according to Alves et al. (1998).Conidial viability was verified under the optical microscope, by means of the percentages of germinated and non-germinated conidia 24 h after plating on PDA + antibiotic and incubation for 24 h in a B.O.D. chamber at the same temperature and photophase previously mentioned.

Selection of the most pathogenic isolates
Ten M. anisopliae and 10 B. bassiana isolates were used (Table 1).One experiment was installed for each fungus, in a completely randomized design, consisting of the isolates and one control with 5 replicates; each plot consisted of 20 mites, totaling 100 mites per treatment.The suspensions were prepared by adding 15 mL of sterilized distilled water plus Tween 80 at 0.01% to the conidia.The suspensions were filtered through sterilized gauze; counts were made in a Neubauer chamber, and the suspensions were standardized at 10 8 conidia mL -1 .Cassava leaf discs 3.5 cm in diameter were inoculated with the pathogen by immersion in 20 mL of the suspension for 5 seconds.The discs were then placed in Petri dishes containing two layers of filter paper circles, moistened with distilled water and left to dry for twenty minutes in a laminar flow chamber.Adult mites aged 0-24 h after emergence from the stock colony were transferred to the arenas with a brush; twenty females per leaf were placed on the abaxial surface of the leaf.The females were sexed visually, by observation of the shape of the opisthosoma, which is round in females but different in males, that have funnel-shaped opisthosomas.The discs were placed and centralized onto polyethylene dishes 9 cm in diameter and 1.5 cm tall with a pin previously attached to the dish with silicone glue; the dishes contained 30 mL distilled water, which served as a physical barrier for the mites and to maintain leaf disc turgidity; the water level was completed daily.
Leaves were replaced on the fifth day after installation to ensure the quality of the substrate, and evaluations were carried out for eight days.The dead mites were placed in a wet chamber to confirm the causal agent and to determine confirmed mortality.
The daily mortality values were accumulated during the experiments to allow LT 50 calculation, by Probit analysis, using the Mobae computer program (Haddad et al., 1995).The mite mortality data were submitted to analysis of variance using F test and the means were compared by Tukey test (P ≤ 0.05) using the Sanest (version 3.0) software package.

Lethal Concentration Estimation
The 645 B. bassiana and CG 321 M. anisopliae isolates, identified as the most pathogenic in the previous stage based on mortality and smaller LT 50 observed against the cassava green mite, were used at the concentrations of 1 × 10 4 , 10 5 , 10 6 , 10 7 , and 10 8 conidia mL -1 for LC 50 determination.The experiments were conducted as a completely randomized experimental design, consisting of six treatments and five replicates containing 20 M. tanajoa females each, totaling 100 mites per treatment.The same methodology employed in the isolate selection experiment was used for disc and suspension preparation, as well as for suspension application, conidial viability determination, mite transfers, and evaluations.

Selection of the most pathogenic isolates
Conidia viability for the M. anisopliae and B. bassiana isolates was higher than 95%.The total mortality percentage caused by B. bassiana isolates varied from 13.0 to 97.0%; confirmed mortality varied from 9.0 to 91.0%, and LT 50 values ranged from 4.2 to 17.0 days.Isolate 645 caused the highest total mortality (97.0 %) and confirmed mortality percentages (91.0 %), and the  2).These values are similar to those obtained by Tamai (1997) for the mite Tetranychus urticae Koch, with a variation from 5.5 to 100% in total mortality and from 4.2 to 73.3% in confirmed mortality, using Beauveria spp.isolates at a concentration of 5 × 10 8 conidia mL -1 .Also, Oliveira et al. (2002), working with B. bassiana isolates at 10 8 conidia mL -1 and the red mite Oligonychus yothersi (McGregor), verified a variation in total mortality from 77 to 98% and a confirmed mortality from 19 to 75%.
With regard to M. anisopliae, the total mortality percentage obtained for isolates of this fungus ranged from 12.0 to 45.0%, confirmed mortality ranged from 8.0 to 45.0%, and LT 50 ranged from 8.6 to 18.4 days (Table 3).When these values are compared to those obtained for B. bassiana, M. anisopliae was less pathogenic to Mononychellus tanajoa; the M. anisopliae isolate CG 321 was the most efficient, providing a 45.0% confirmed mite mortality with an LT 50 of 8.6 days.These results are better than those obtained by Tamai (1997), who evaluated M. anisopliae against T. urticae; the author obtained low pathogenicity against the mite, achieving only 4.2 % confirmed mortality.
Even though using a concentration of 2.27 × 10 6 conidia mL -1 , Albuquerque et al. (2000), during an evaluation of M. anisopliae pathogenicity over Brevipalpus   phoenicis Geijskes, also did not verify efficiency of the fungus over this mite; during their study, only 2.25% mortality occurred, and no conidiogenesis was observed.

Lethal Concentration Estimation
A relationship was verified between the increase in conidial suspension concentration and the increase in total mortality and confirmed mortality caused by the B. bassiana 645 and M. anisopliae CG 321 isolates, as well as with the LC 50 values (Table 4).The total mortality caused by the B. bassiana isolate 645 at the concentration of 10 4 conidia/mL was not different from the control; however, the other concentrations caused higher mortalities than the control, with a total mortality of 93.0% and a confirmed mortality of 83.0%, at the concentration of 10 8 conidia mL -1 .
The LC 50 for the 645 B. bassiana isolate was 3.93 × 10 6 conidia mL -1 , ranging from 2.67 × 10 4 to 7.36 × 10 7 conidia mL -1 (Table 4).The 645 B. bassiana isolate showed a lower LC 50 than the CG321 M. anisopliae isolate, thus demonstrating that the first was more pathogenic than the second.Tamai et al. (1999), using the 447 B. bassiana isolate on females of the mite T. urticae, concluded that only at the concentration of 1 × 10 9 conidia mL -1 there was a total mortality higher than 50%.Alves et al. (2002), evaluating the effect of B. bassiana isolate 447 on T. urticae at different concentrations, verified an LC 50 of 1.26 × 10 7 conidia mL -1 , therefore higher than the value obtained in this work with B. bassiana isolate 645.
With regard to the performance variability of different isolates, Sosa Gómez & Alves (1983) verified a high enzymatic activity in more virulent isolates of M. anisopliae from several Brazilian regions, and mentioned that they are probably associated with the presence of enzymes that influence the penetration process of the fungus (St Leger et al., 1988;De La Rosa et al., 1997), as well as with toxins such as destruxins and beauvericin present, respectively, in M. anisopliae and B. bassiana, which vary in different isolates (Roberts & Krasnoff, 1998).Considering that the use of entomopathogenic fungi should be viewed as a component of integrated pest management, the results here obtained suggest that B. bassiana isolate 645 should be used in control programs against the cassava green mite M. tanajoa.26.0 ± 6.5 ab 19.0 ± 5.3 ab (1.9 x 10 8 -6.5 x 10 9 ) 10 6 21.0 ± 5.5 bc 13.0 ± 3.9 bc 10 5 12.0 ± 2.5 bc 7.00 ± 1.

Table 1 -
Hosts and origins of Metarhizium anisopliae and Beauveria bassiana isolates used on Mononychellus tanajoa in laboratory experiments.
1 Means (± SE) followed by the same letter in the columns do not differ by Tukey test at 5%.2LT 50 Lethal Time; CI, Confidence Interval at 5%; CV, Coefficient of Variation.
1 Means (± SE) followed by the same letter in the columns do not differ by Tukey test at 5%. 2 LC 50 , Lethal Concentration; CI, Confidence Interval at 5%; CV, Coefficient of Variation. 3b Regression line slope coefficient 4 χ 2 = χ 2 Test . bassiana isolates tested are more efficient than the M. anisopliae isolates for utilization in integrated management of the cassava green mite M. tanajoa.