BACKGROUND: Gaseous ozone (O3) has potential for control of insects in stored grain. Previous studies have focused on freely exposed insects. Immatures of internal pests (e.g. Sitophilus spp. and most stages of Rhyzopertha dominica F.) are protected within kernels and probably require higher doses and/or longer treatment times for full control. A laboratory study determined the doses of ozone necessary for full control of freely exposed and internal stages of eleven stored-product pest species. Test insects were three species of Sitophilus, R. dominica, Tribolium confusum Jacquelin du Val, T. castaneum Herbst, Plodia interpunctella H¨ ubner, Sitotroga cerealella Olivier, Oryzaephilus surinamensis L., Ephestia kuehniella Zeller and Stegobium paniceum L. Insects were exposed to continuous flows of ozone in doses of 10–135 ppm and exposure times of 5–8 days. Dose-mortality bioassays were conducted on three species of Sitophilus and P. interpunctella.

RESULTS: Freely exposed stages (with a few exceptions) were controlled with 35 ppm of ozone for 6 days. Full mortality of internal stages within kernels required exposure to 135 ppm for 8 days.

CONCLUSION: This study confirms that higher doses and/or longer treatment times are necessary for control of internal stages of stored-product pests.

1. INTRODUCTION Grain stores are a perfect habitat for insects that live on dry seeds: the insects are protected from weather extremes, have access to an unlimited food resource and are undisturbed for long periods. Insect infestation in stored grain can result in economic losses of up to 9% in developed countries and 20%, locally much higher, in developing countries. Furthermore, the presence of insects and their remains in grain may pose a health risk to humans and livestock.

At present, pests in commercial grain stores are managed by a combination of different methods ranging from cleaning and cooling to treatment of the grain with contact insecticides, or fumigation. Some of the economically most important pests of stored grain worldwide have stages that feed and develop concealed within kernels, i.e. eggs, larvae and pupae of the weevils Sitophilus granarius L., S. oryzae L. and S. zeamais Motschulsky (Coleoptera: Curculionidae) and older larvae and pupae of the lesser grain borer, Rhyzopertha dominica F. (Coleoptera: Bostrychidae). Internal stages are not directly affected by contact insecticides. Fumigation with phosphine has an effect on internal stages but is problematicbecauseof itshighacute toxicity tohumansandability to provoke development of resistantpest strains. The available pesticides for treatment of grain are decreasing owing in some cases to environmental and safety concerns among consumers and society, thus emphasising the need for alternative control methods.

Gaseous ozone (O3) is highly oxidative and unstable and decomposes rapidly to oxygen without leaving residues. It is a powerful disinfectantused for water treatment andin the food industry and has recently received increasing interest for control of insect pests in stored grain. Several papers report on the potential of ozone to control stored-product pests. Ozone has also shown potential for control of insect strains that are resistant to phosphine. Lethal doses and exposure times for ozone have been reported to range between 5 parts per million by volume (ppm) for 5 days to 300 ppm for 18 h for insects living freely among kernels.

In previous studies, unprotected specimens of the insects have been used, i.e. adults or larvae that live freely among kernels or in flour, which below are referred to as ‘freely exposed stages’. To the present authors’ knowledge, no studies have been conducted to determine the lethal doses necessary to kill immature stages of internally feeding pests, which below are referred to as ‘internal stages’. They are to some extent protected from exposure to ozone, and it is highly probable that higher doses and/or longer treatment times are necessary to eradicate them. As these internally feeding pests are widespread in grain stores all over the world, it is imperative to determine the lethal doses for the internal stages as well if this method is to be viable for control of pests in stored grain and other stored products.

The present paper sets out the results of a laboratory study to determine the lethal doses of ozone needed to control adults and immatures of eleven species of stored-product pests: eight species of Coleoptera – Oryzaephilus surinamensis L., Rhyzopertha dominica, Sitophilusgranarius, S. oryzae, S. zeamais, Stegobium paniceum L., Tribolium castaneum Herbst and T. confusum Jacquelin du Val, and three species of Lepidoptera: Plodia interpunctella H¨ ubner, Ephestia kuehniella Zeller and Sitotroga cerealella Olivier. All stages of each species were used in the investigation, a total of 9172 treated specimens and an equivalent number as untreated controls. Dose–mortality bioassays were conducted on the three species of Sitophilus and P. interpunctella. The lethal doses found for these species were then tested on the remaining seven species. The insects were exposed to continuous flows of ozone in doses of 10 to 135 ppm and exposures times of up to 8 days.

2. MATERIALS AND METHODS 2.1 Insects

The insect species and stages used in the study and their rearing conditions are listed in Table 1. New colonies were established weekly by transferring adults to rearing media specific for each species. The parent insectswere removed after 2 days (S. Granarius and S. oryzae after 1 day), and containers with offspring were incubated under the rearing conditions until testing. Internal stages were treated while concealed within kernels or other products. Details of the age groups of the test specimens are given in Table 1.

The insects were exposed to ozone in shallow plastic containers (diameter 65 mm, height 22 mm; Færch Plast, Holstebro, Denmark) with ventilated lids. Eggs of external feeders were transferred by means of a moist brush to black filter paper in the container. In the case of freely exposed stages, specimens were transferred to the container along with a very small amount (∼0.5 mL) of dry rearing medium. In most of the tests with freely exposed stages, at least ten specimens of each stage were used in each replicate. For tests with internal stages, a number of infested kernels (in the case of S. paniceum, suitable amounts of rearing medium) were taken to obtain a suitable number of test specimens (15–30 or more).As the kernels were picked at random from the same source, the numbers of test specimens were considered to be equal in treated and untreated units. After the initial pilot trials, at least three replicates in time were conducted with each species and stage. In all trials the same numbers of specimens were used as untreated controls.

2.2 Ozonation The treatments were carried out in controlled climate chambers at 20±1 ◦C and 65±5% RH. The ozone was generated by a custommade ozone generator using the corona discharge method. The temperature of the air leaving the generator was regulated by leading it through a pipe in a water bath. A pump pulled the gaseous ozone through a perspex box (length 115 cm, width 60 cm, height 20 cm) in which containers with test insects were placed on a grate 2 cm above the bottom to ensure an even distribution of the ozone. The untreated controls were placed in a similar box in a similar chamber under the same climatic conditions for the same duration.

The ozone level in the treatment box was monitored online (ozone monitor UV-100; Eco Sensors Inc., Santa Fe, NM). Extra checks of ozone levels were conducted with gas detector tubes (range 2.5–100 ppm, Kitagawa No. 182SB; Komyo Rikagaku Kogyo KK, Japan), which were also used as the sole method for ozone measurement in a few tests when the online monitor was out of order after extended exposure to ozone (five out of a total of 37 ozone treatments).

After exposure to ozone, test specimens were incubated under colony maintenance conditions (Table 1). Freely exposed stages were observed daily for a minimum of 5 out of 7 days up to a maximum of 7 days or until all the treated insects had died. Containers with internal stages were kept until the specimens in the untreated controls had emerged as adults, a minimum of 3 weeks.

Arange of different combinations of ozone dosage and exposure time were tested. Initial pilot tests were conducted to define the lethal combination of dosage and exposure time for an external feeder (Plodia interpunctella) and internal feeders (three species of Sitophilus) (Table 1). These tests started at very low dosages (10–20 ppm for 5 days). Dosages and treatment times were increased in an attempt to obtain full mortality of all stages. Tests were then conducted at approximately 35, 75, 100 and 135 ppm with exposure times of 5, 6 and 8 days.

When full mortality in these four species was obtained, tests were conducted on the remaining seven species with two different treatment regimes: one for freely exposed stages (35 ppm, 6 days), and one for the internal stages (135 ppm, 8 days). A few other combinations were tested on selected species and stages.

2.3 Data analysis Data from tests with internal stages of Sitophilus granarius and S. oryzae exposed to the same exposure time (5 days) were suitable for probit analysis. Dose–mortality responses were analysed by probit analysis and LD values estimated. When χ2 was small (P > 0.10), confidence limits were calculated using t = 1.96. When χ2 was high (P < 0.10), heterogeneity was suspected, and t for calculationof confidence limitswasbasedonthe heterogeneity factor.

The datasets from tests with P. interpunctella and Sitophilus zeamais were not suitable for probit analysis. The lethal doses derived from the tests with P. interpunctella and Sitophilus spp. proved to be lethal for the remaining species in most cases, and data analysis was inapplicable.

3. RESULTS During each test, the ozone level was very stable; the standard deviation in 75% of the tests was less than 4.6% of the mean. However, small differences occurred between the mean ozone levels among tests, even though they were set at the same conditions and settings. Survival in the untreated controls of tests with freely exposed stages was above 85% in the great majority of the tests (159 out of 182 tests), and generally higher than 95%.

As regards the internal stages, the average number of specimens in the control kernels for S. granarius was 21.3 ± 6.2 (± SD, 131 tests) and for S. oryzae it was 31.1 ± 10.1 (53 tests). The variation among tests is considered to be acceptable considering the challenges involved with obtaining uniform infestation rates with internal pests, e.g. inherent biological differences in the parent material, the grain and the development rate of the insects. With S. zeamais, only low infestation levels were obtained, and in some tests the number of replicates was increased to compensate for this. However, the results with this species were not as soundly based as the rest.

In the tables, the average of the ozone doses of the replicates for each species and stage are given, but for simplicity the dosage averages of 20, 35, 75, 100 and 135 ppm are used in the presentation of results and the discussion.

In the great majority of tests with freely exposed specimens, 100% mortality was obtained at 35 ppm for 6 days (results from treated specimens in Table 2a; untreated controls in Table 2b). This applies to adults of the tested pest species, immature stages of T. castaneum, T. confusum, O. surinamensis, P. interpunctella and E. kuehniella and eggs of R. dominica, S. cerealella and S. paniceum. A few exceptions were found: 15% of S. paniceum adults survived this treatment. The resultswith eggs of P. interpunctella were more variable, and slight survival was observed at 75 ppm for 5 days as well as 20% survival at 35 ppm for 6 days. Adults of several species were tested at 20 ppm for 5 days, resulting in 100% mortality (E. kuehniella, O. surinamensis (no replications), P. interpunctella, S. cerealella, Sitophilus spp., T. confusum). An extra test revealed that 100% mortality in adults of all three species of Sitophilus is obtained after 24 h at 135 ppm (three replicates, ten specimens per replicate, full survival when untreated).

As for the internal stages, S. oryzae larvae were more tolerant to ozone than the other two species of Sitophilus (Table 3). In medium-aged larvae and pupae of this species, a small survival was observed at 75 and 100 ppm, both for 5 days, and for young larvae even at 135 ppm for 8 days (1% survival). All stages of S. zeamais and all of S. granarius except eggs are controlled when exposed to 100 ppm for 5 days. In subsequent tests the dosage of 135 ppmfor 8 days proved to be lethal for the internal stages of the other species tested, i.e. R. dominica, S. cerealella and S. paniceum.

The results of probit analysis of data from immature stages of S. granarius and S. oryzae treated for 5 days are shown in Table 4. In both species the slope of theprobit lines decreasedwith increasing age of the test specimens. Heterogeneity was suspected in the results with S. granarius. Test specimens in different stages were selected on the basis of age of the colony, and not on direct observation, as the larvae or pupaewere located inside the kernels. Thus, some specimens in a given samplemay have developed into the next stage,and this may have contributed to the heterogeneity. When heterogeneity occurs, the fiducial limits in the upper and lower ends of the mortality curve are very broad. Comparison between groups should therefore be done around LD50. LD50 values of both species ranged from 5 to 27 ppm; if the same stages were compared, they were higher in S. oryzae than in S. granarius. The LD95 values ranged from 29 to 60 ppm in S. granarius and from 61 to 100 ppm in S. oryzae, but with extremely wide fiducial limits.

4. DISCUSSION AND CONCLUSIONS The present investigation was designed to provide general guidelines for the dosage of ozone that is required to control a long list of stored-product pest species, including the internal stages that are protected within kernels. A treatment that only controls the freely exposed stages will not maintain the quality of the grain, as the insects within kernels will emerge as adults and establish a new generation of pests. Information on lethal doses for internal stages will therefore increase the applicability of ozonation as a control method in grain stores.

The most thorough testing was conducted with three species of Sitophilus. To be able to obtain full control of all internal stages of these pests, a treatment with 135 ppm for 8 days is necessary. This dosage is recommended for all three Sitophilus species for all practical reasons, in spite of the very slight survival seen at this dose in small larvae of S. oryzae (one specimen compared with 92 in the untreated controls). The present study did not include tests between 100 and 135 ppm or exposure times between 6 and 8 days, but the results with S. oryzae indicate that these intermediate doses or exposure times will probably not be efficient.

As shown in the probit analysis of data obtained with immature S. granarius and S. oryzae exposed for 5 days (Table 4), there is a decrease in the slope of the probit line from stage to stage in both species. This indicates that young larvae have a critical dose at which they are all killed. Greater diversity is seen in the response of older larvae and/or pupae. Even at low doses, some mortality is seen in all stages, but there is clearly a robust group of insects that can survive larger amounts of ozone. Some robustness may stem from the fact that old larvae and pupae are physically large. It may also be presumed that the amount of endosperm that has been consumed by the larva and the distance of the larva or pupa from the seed coat have an influence on the effect of the ozone. Kernels infested by Sitophilus spp. suffer a significant weight loss; larvae of S. granarius each consume more than 60% of a wheat kernel. Especially for S. granarius, an indication of this can be seen when the LD50 values are compared: the sensitivity to ozone increases between the second and third larval/pupal stage, indicating that, even though the latter is larger and presumably more robust, it is more exposed to the ozone, as there is hardly any contents left in the grain, leaving the seed coat as the only protection.

With S. oryzae it was not possible to run a probit analysis on the data from young larvae treated for 5 days, as all specimens survived at the lowest dose (20 ppm) and all died at the next dose (75 ppm), suggesting a very steep slope. Even so, as with S. granarius, S. oryzae seemed to exhibit a somewhat larger spread in sensitivity as the larvae grew and changed stages. Furthermore it seems that larger larvae and pupae are more robust. With S. oryzae, the sensitivity did not decrease at high doses, and LD50 was approximately 5 times higher than LD50 for S. granarius. S. oryzae thus seems to be more robust than S. granarius, but it has to be emphasised that the heterogeneity in the experiments with S. granarius affects the interpretation of the results. Some variation also exists in S. oryzae, as the single surviving larva at 135 ppm for 8 days suggests.

Freely exposed stages of the eleven species proved to be more susceptible to ozone, and, except for P. interpunctella eggs, full mortality was generally obtained at 35 ppm for 6 days, for some species and stages even less. Several studies have reported on lethal dosages and exposure times that are much lower than the ones found in the present study. Full mortality was obtained in T. castaneum and T. confusum adults after exposure to 45 ppm for 3.5 to 6.5 h, depending on stage. This is much shorter than the results of another study in the present project into the effects of sublethal doses of ozone on gene expression and lipid peroxidation.17 Here, larvae and adults of T. castaneum were exposed to 40 ppm of ozone in stepwise intervals of 6 h. Larvae proved to be more susceptible to ozone than adults: in larvae full mortality occurred after 24 h at 40 ppm, whereas adults needed 48 h exposure for close to full mortality. In another study, 100% mortality was found in adult T. confusum and O. surinamensis after 5 and 3 days exposure to 5 ppm respectively. In field experiments in grain stores, an ozone dose of 25 ppm for 2 days led to ∼100% mortality in adult S. oryzae, as did 50 ppm for 4 days in T. castaneum. However, in adults of four other species of storedproduct pests, as well as in immatures of P. interpunctella, less than 100% mortality was found after treatment for 4 days.12 Until the reasons for these differences are clarified, it is suggested that doses of 35 ppm for 6 days be used for freely exposed pests.

The present study shows that ozonation has potential as an efficient control method against most insect pests in stored products, thus confirming previous sporadic studies that focused on only a few insect species and mainly adult insects. For the freely exposed stages of the eleven species, full control can generally be obtained with 35 ppm for 6 days. However, for control of internal stages of Sitophilus spp. and R. dominica, full mortality requires approximately 135 ppm for 8 days. This seems to be the first report of lethal doses of ozone on internal stages of stored-product pests. These tests were conducted with the internal stages in wheat; the effect in other species of grain, e.g. maize, needs to be tested for practical recommendations for the use of ozone in these commodities. Treatment of a grain store with these high doses of ozone requires constant monitoring and the occasional addition of ozone. However, it seems that, in the first phase of ozonation, the ozone is rapidly degraded. After the molecular sites that degrade the ozone are saturated, very little degradation occurs. Thus, it is probable that a high level of ozone can be maintained, but this needs further investigation.

Ozone is attractive owing to the fact that (i) it can be generated on location, (ii) it leaves no toxic residues and (iii) risks to operating staff are minimised owing to its rapid decomposition. Ozonation thus has the potential to become a realistic choice for control of pests in stored products either alone or as a supplement to other methods.

ACKNOWLEDGEMENTS Lars Damberg, Bodil M Pedersen and Jørgen Christensen are thanked for technical assistance. Two anonymous reviewers are thanked for valuable comments. This project was supported by the Danish Ministry of Food, Agriculture and Fisheries, Aarhus University and Crop-Protector KS.