NYMPHAL DEVELOPMENT AND FEEDING PREFERENCE OF Podisus maculiventris ( HETEROPTERA : PENTATOMIDAE ) ON EGGS OF Ephestia kuehniella ( LEPIDOPTERA : PYRALIDAE ) PARASITISED OR NOT BY Trichogramma brassicae ( HYMENOPTERA : TRICHOGRAMMATIDAE )

Predation by Podisus maculiventris nymphs, a predatory pentatomid, was evaluated with eggs of the flour moth Ephestia kuehniella (Pyralidae), parasitised or not by Trichogramma brassicae (pupae stage). Eggs of this pyralid were glued on rectangular cardboard and presented to nymphs of P. maculiventris as food. The pentatomid successfully reached adult stage when feeding on unparasitised eggs, indicating that flour moth eggs can be used as a factitious food for rearing this predator. Pentatomid nymphs that received only parasitised eggs died before reaching fourth instar. In choice tests, P. maculiventris showed a preference for preying on unparasitised eggs of E. kuehniella rather than those containing pupae of T. brassicae. These results show that it is possible to combine the use of P. maculiventris with releases of T. brassicae in control programs of lepidopteran pests.


INTRODUCTION
Pentatomid predators including species of the genus Podisus are natural enemies of many insect pests in North America (McPherson, 1980;De Clercq, 2000), Costa Rica, Panama, Peru and Bolivia, Argentina (Thomas, 1992), and Brazil (Zanuncio et al., 1994(Zanuncio et al., , 2001;;Torres & Zanuncio, 2001).In South America, the beneficial action of these natural enemies has been reported in plantations of Eucalyptus (Zanuncio et al., 2000(Zanuncio et al., , 2001) ) and in many agricultural crops such as coffee (Gravena & Lara, 1982), and cashew (Silva, 1965), and passion fruit (Costa Lima, 1940).Predatory Pentatomidae are important agents in biological control of defoliating caterpillars and other insects either through natural occurrence or in programmed releases (Zanuncio et al., 1994;De Clercq, 2000).However, success of biological control depends on a good knowledge of the interaction between predatory Pentatomidae and other natural enemies present in a crop.
Generalist predators like Orius, Geocoris, and Nabis can attack lepidopteran eggs parasitised by Trichogramma spp. with losses up to 50% in corn and 91% to 98% in cotton (Smith, 1996), but predation by Podisus on parasitised eggs has not yet been reported.Anthocorid predators are more likely to accept unparasitised host eggs than those containing pupae of Trichogramma, although younger stages of Trichogramma are equally susceptible (Brower & Press, 1988).Trichogramma and Podisus may have potential for being used simultaneously in biocontrol programs against lepidopteran pests but interactions between these natural enemies need to be addressed.The objective of the current research was to study feeding preference, development rate, and survival of nymphs of P. maculiventris reared with eggs of the flour moth Ephestia kuehniella (Zeller, 1879) (Lepidoptera: Pyralidae), parasitised or not by the parasitoid Trichogramma brassicae (Bezdenko, 1968) (Hymenoptera: Trichogrammatidae).Since Trichogramma have been released in host eggs containing pupae near hatching (Pinto & Parra, 2002), we used seven-to eight-day-old eggs of E. kuehniella with pupae parasitoids.

MATERIALS AND METHODS
This study was carried out in the Laboratory of Agrozoology, Department of Crop Protection, Ghent University, Belgium.All experiments were done at 23 ± 1ºC, relative humidity of 75 ± 5%, and 16 h photophase.Nymphs of P. maculiventris were taken from a colony maintained in the Laboratory of Agrozoology; eggs of E. kuehniella parasitised or not by T. brassicae were obtained from Koppert Biological Systems, Berkel en Rodenrijs, The Netherlands.Parasitised eggs of Ephestia containing seven to eightday-old parasitoids (i.e., pupae) were kept in a refrigerator at 4-6°C for a maximum of seven days to avoid further development of the parasitoid.
In a first experiment, one-day-old second instar nymphs of P. maculiventris were placed in individual 9.2 x 1.5 cm Petri dishes with a slice of green bean for water supply.In the first treatment, 10 nymphs of the predator were fed ad libitum with eggs of E. kuehniella parasitised by T. brassicae; in the second treatment, 10 nymphs received unparasitised eggs of the pyralid.Lepidopteran eggs were glued on rectangular cardboards and fresh eggs were provided every 24 hours as needed.Duration of each instar and survival of P. maculiventris were monitored.
In a second experiment, 10 nymphs of P. maculiventris, in the first day of the second instar, were placed in 5.0 x 1.8 cm Petri dishes and kept during 24 hours without prey and with a slice of green bean for water supply.After this period, nymphs of this predator received, at the same time, two white cardboards of similar size (2.0 x 0.5 cm).The first cardboard had glued unparasitised eggs of E. kuehniella; the other contained eggs of E. kuehniella parasitised by T. brassicae.Observations were made up to 120 minutes after starting the experiment to evaluate preference of P. maculiventris feeding.
An olphactometer was not used in these tests because this methodology does not permit the evaluation of feeding change on parasitised and unparasitised eggs of E. kuehniella by P. maculiventris over time.
Data from both experiments were submitted to analysis of variance; means were compared with the test of Wilcoxon with a 5% probability level.

RESULTS
The duration of the second instar of nymphs of P. maculiventris in the first experiment (4.4 days) was significantly shorter when they were fed with unparasitised than parasitised eggs of E. kuehniella (6.0 days); it was 4.90, 4.25, and 6.62 days for the third, fourth, and fifth instars, respectively, when fed on unparasitised eggs (Table 1).Survival of P. maculiventris in the second instar was similar with eggs of E. kuehniella parasitised or not by T. brassicae.However, all nymphs of P. maculiventris fed on parasitised eggs died during the third instar.
In the second experiment, nymphs of P. maculiventris did not show any preference at the first attack for parasitised or unparasitised eggs of E. kuehniella glued on the rectangular cardboard.However, nymphs of this predator were seen in about 80% of the cases feeding on unparasitised eggs from 15 minutes after the start and up to the end of the experiment (Table 2).Means at the same column followed by the same letter do not differ by the test of Wilcoxon at 5% probability level.

Second instar Third instar Fourth instar Fifth instar
-All nymphs died during the third instar.

DISCUSSION
The development period from second instar to adulthood of P. maculiventris with unparasitised eggs of E. kuehniella (20.2 days) was similar to that reported for this predator when fed with larvae of Galleria mellonella (L., 1758) (Lepidoptera: Pyralidae) (De Clercq & Degheele, 1992) and S. exigua (De Clercq & Degheele, 1994), which are considered good prey for this predator.Furthermore, 80% of nymphs successfully reached adult stage when exclusively fed on unparasitised E. kuehniella eggs.These findings suggest that flour moth eggs can be used as a factitious food to rear P. maculiventris.
The duration of second instar was prolonged by 36% when nymphs of P. maculiventris were fed with parasitised flour moth eggs, reflecting the poor quality of this diet.Nymphs of this predator showed high survival in the second instar with parasitised eggs of E. kuheniella but all of them died during third instar.This indicates that eggs with T. brassicae pupae do not constitute an adequate food for P. maculiventris.
In a choice situation, second-instar nymphs of P. maculiventris did not show a clear preference for either parasitised or unparasitised eggs at the first attack, but a significant difference was found after fifteen minutes when most predators were observed feeding on unparasitised eggs.Thereafter, predators were seen feeding on unparasitised eggs in 80% of the cases.This finding suggests that parasitised eggs of E. kuehniella were rejected by P. maculivetris after initial feeding, a preference which may be due to the fact that eggs of this Lepidoptera with pupa of Trichogramma have reduced acceptability by this predator.The predator Xylocoris flavipes (Reuter, 1875) (Heteroptera: Anthocoridae) preferred unparasitised host eggs to those containing pupae of the parasitoid Trichogramma pretiosum Riley, 1879, but host eggs containing parasitoids in earlier development phases were as susceptible to predation as unparasitised eggs (Brower & Press, 1988).Orius insidiosus (Say, 1832) (Heteroptera: Anthocoridae) showed a similar attack rate on parasitised or unparasitised eggs of H. virescens and degree of parasitization by T. pretiosum through consumption of their eggs was directly influenced by population density of this predator (Lingren & Wolfenbarger, 1976).Thus, the outcome of interaction between P. maculiventris and egg parasitoids of the genus Trichogramma may depend on developmental stage of the parasitoid.
It is important to understand predation by arthropod predators on host eggs parasitised by Trichogramma spp.because loss of these eggs in augmentative releases of Trichogramma may reach 50% in corn (Yu & Byers, 1994) and from 91% to 98% in cotton (Jones et al., 1977).The findings of the current study suggest that P. maculiventris can be used in combination with releases of parasitoid wasps of the genus Trichogramma provided that the eggs of the lepidopteran host contain pupae of the parasitoid.However, further evaluation is recommended of feeding behavior of this and other predators when presented with eggs of different hosts containing parasitoids of different ages.

TABLE 1 Development time in days (D) and percentage of nymph survival (S) of Podisus maculiventris fed on eggs of Ephestia kuehniella parasitised (T1) or not (T2) by Trichogramma brassicae at 23 ± 1ºC, relative humidity of 75 ± 5% and a 16 h photophase. TABLE 2 Predation by Podisus maculiventris on eggs of Ephestia kuehniella parasitised (T1) or not (T2) by Trichogramma brassicae in a choice test at 23 ± 1ºC, relative humidity of 75 ± 5% and 16 h photophase.
Means within a row followed by the same letter do not differ by the test of Wilcoxon at 5% probability level.