The relative susceptibility of life stages may also vary among pesticides

To determine acute mortality via oral exposure, A. pisum were treated by being individually dipped in a particular pesticide solution. They were removed from the solution with a fine paint brush, and placed on a paper towel in a fume hood until excess pesticide had dripped off and the aphids had dried sufficiently and were able to crawl. 10 topically-treated adult A. pisum were placed in a test arena, and a single untreated H. convergenswas added and allowed to feed for a period of 48 h. To test the effects of residual exposure, individual arenas were treated by adding 4 ml of a given pesticide solution and turning 360 degrees in order to coat the entire inside surface before the solution was poured out. Vials were inverted, suspended on a rack, and allowed to dry in a fume hood for a period of 24 h. Approximately 10 untreated adult A. pisum were then added as a food source to the treated arenas. A single untreated H. convergenswas added per arena and allowed to crawl on the treated surface for a period of 48 h. For topical exposure,hydroponic fodder system individual H. convergens were placed in an 8.5 cm petri dish with filter paper lining and treated in a Potter spray tower set to 68.9 kPa. For each pesticide solution, the spray volume per application was 1.4 ml, resulting in a spray deposit of 2.50 mg cm-2 .

This deposit is similar to that used in other studies on non- target insects as well as by the IOBC Working Group “Pesticides and Beneficial Organisms” . The treated insects were allowed to dry, and were placed individually into an untreated arena with 10 untreated A. pisum adults as a food source for a period of 48 h. The statistical program R was used for all analyses. Generalized linear models with binomial errors, or quasibinomial errors in cases where there was over dispersion, were used to analyze the effects of pesticide, pesticide concentration , life stage , and exposure route on the acute mortality of H. convergens. For all GLMs, model reduction and log likelihood ratio tests were used to assess the statistical significance of each factor and their interactions on acute mortality at α = 0.05. In a first analysis, the uncorrected acute mortality data were analyzed. Starting with a full model that included all four factors and interactions, the data set was finally subdivided into eight separate models with no significant interaction between factors to compare the relative toxicity of the different pesticides tested. From these models, the pesticides and pesticide concentrations with similar effects on acute mortality were grouped , and their toxicity was characterized according to the criteria developed by the International Organisation for Biological Control for classifying pesticide selectivity on beneficial insects in a laboratory bio-assay: <30% mortality is harmless, 30-79% mortality is slightly harmful, 80-99% mortality is moderately harmful, and mortality > 99% is harmful . Although uncorrected mortality was used in our analysis, the IOBC classification remains applicable due to the fact that control mortalities were low. This classification scheme has been widely adopted for the comparison of differential toxicities of pesticides to natural enemies in a number of laboratory studies .

In a second analysis, all acute mortality data were first corrected to account for variability among control groups , and then used to examine comparative effects of life stage and route of exposure. An initial full model included all four factors, pesticide, pesticide concentration, life stage and exposure route, and their interactions as categorical variables. Due to significant interactions, the corrected data set was then further separated by pesticide and pesticide concentration to examine the effects of life stage and exposure route on acute mortality of H. convergens. Additionally, the interaction for effects between pesticide and gender were tested for adults. Of the seven pesticides tested, only lambda-cyhalothrin and cyantraniliprole were either moderately harmful or harmful, causing greater than 80% mortality. Lambda cyhalothrin was either moderately harmful or harmful to both life stages of H. convergens when topically exposed at the 100% concentration or when residually exposed at either concentration. However, for oral exposure at both concentrations, it was moderately harmful or harmful to larvae, but not to adults. Cyantraniliprole was moderately harmful to larvae, but not to adults at the 100% concentration when orally exposed, but it was harmless or slightly harmful at the 10% concentration and for other exposure routes. Chlorantraniliprole, copper+mancozeb, novaluron, spinetoram, and sulfur were either harmless or slightly harmful to both life stages at both concentrations, regardless of exposure route. The acute toxicities of the seven pesticides to H. convergens in this study were classified according to guidelines initially established by the IOBC for laboratory bio-assays due to the widespread use of this classification.

However, a more recent revision of these guidelines for classifying pesticide selectivity may more accurately describe their potential toxicity to natural enemies in the field . The revision suggests reducing the number of categories from four to three; the harmless category remaining unchanged at less than 30% acute mortality, pesticides responsible for 30-79% acute mortality being changed from slightly to moderately harmful, and the top two categories of 80-99% and > 99% acute mortality being combined into a single harmful category. Although these revised categories have yet to be widely adopted, they offer a valuable improvement over the original classification as LC50 values would no longer be considered only slightly harmful and the artificial separation of > 99% mortality as a distinct category would be dropped. Under these revised guidelines, lambda-cyhalothrin would be considered harmful rather than moderately harmful to larvae for both concentrations and all three routes of exposure, and to adults at both concentrations for residual exposure and at the 100% concentration for topical exposure. Additionally, lambda-cyhalothrin would change from slightly harmful to moderately harmful for adults orally exposed to the 100% concentration and for adults topically exposed to the 10% concentration. Cyantraniliprole would also be reclassified from slightly harmful to moderately harmful at the 100% concentration for adults orally exposed and for larvae topically exposed, and from moderately harmful to harmful for larvae orally exposed to the 100% concentration. Spinetoram and novaluron would also change from slightly harmful to moderately harmful for topical exposure of larvae at the 100% concentration. Copper+mancozeb and sulfur would remain in the harmless category for all life stages, concentrations, and exposure routes. All further discussion of toxicity will be used in reference to the revised classification system. For corrected acute mortality, life stage had no significant effect for chlorantraniliprole, cyantraniliprole, spinetoram and sulfur at the 100% concentration or for copper+mancozeb at both concentrations. The most pronounced life stage effects were for cyantraniliprole, lambdacyhalothrin, novaluron, and spinetoram at the 100% concentration, and with larvae being more susceptible than adults in all cases. Similarly, novaluron has been shown to have a significant acute effect on larvae of C. carnea, but not on adults . In contrast, while sulfur had limited effects on both adults and larvae of H. convergens in this study,hydroponic nft gully for G. occidentalis, it had significant effects on larvae while being harmless to adults . However, immature life stages are not always more susceptible than adults to pesticide exposure. For instance, phosmet and dimethoate were harmless to eggs and pupae of C. carnea, but harmful to adults .

For instance, spinosad was harmless to both adults and nymphs of the predatory mirid bugs Macrolophus pygmaeusand Nesidiocoris tenuis, while indoxacarb was harmless to M. pygmaeus nymphs, but moderately harmful to adults . Though testing a single life stage can be a good starting point for assessing pesticide toxicity, it is important to identify the most susceptible life stage. For H. convergens, larvae were more susceptible to pesticide exposure than adults at the 100% concentrations, but not necessarily for the more dilute 10% concentrations. A similar pattern was found for exposure route; effects were more frequent at the 100% concentration, for larvae than for adults, and for cyantraniliprole, novaluron, lambdacyhalothrin and spinetoram rather than the other pesticides tested. For cyantraniliprole at the 100% concentration, oral exposure caused the greatest mortality for both larvae and adults, whereas for the other pesticides, when there were significant effects of exposure route, topical exposure caused the greatest mortality. Lambda-cyhalothrin proved to be an exception to this pattern with residual exposure having the greatest effect on adults at the 10% concentration, residual and topical exposure having a similar effect for adults at the 100% concentration, and all exposure routes having a similar effect for larvae at both concentrations. Other studies have also found that exposure route can affect mortality rates of natural enemies. For instance, the effects of spinetoram on adult females of the parasitoid Neochrysocharis formosawere classified differently for all exposure routes: harmful via oral exposure, harmless via topical exposure, and moderately harmful via residual exposure . Additionally, lambda-cyhalothrin was harmless to N. formosa females via oral exposure, but moderately harmful via topical or residual exposure. For spirotetramat, topical exposure was moderately harmful to adults of the parasitoid Tamarixia radiata, while residual exposure was harmless . At the maximum label rate, novaluron was harmful to nymphs of Podisus maculiventris Say via residual or oral exposure . In contrast, in our study, novaluron was moderately harmful to H. convergens larvae via topical exposure, but was harmless via residual or oral exposure. Given the variable effects of exposure route on mortality, bio-assays for the effects of reduced risk pesticides on natural enemies should include multiple routes of exposure rather than being restricted to residual exposure alone, the approach that is recommended in the first tier of the IOBC testing guidelines . As testing for all individual exposure routes can be considerably more time consuming, a more practical approach would be to use a combined exposure scenario, where insects are exposed via multiple routes simultaneously rather than each route individually. Another potential benefit to testing multiple exposure routes simultaneously rather than individually is that it would allow for any additive or synergistic effects. assessed the toxicity of reduced-risk insecticides to adults and larvae of two lacewing species, C. carnea and C. johnsoni, by simultaneously treating insects topically, residually, and orally. From measures of survivor ship 10 days after exposure, chlorantraniliprole and cyantraniliprole proved to be harmless to larvae, but harmful to adults. In addition, from the same study, spinetoram and lambda-cyhalothrin were moderately harmful to larvae and harmful to adults. In a similar study of the survivor ship of the predatory mite G. occidentalis from simultaneous routes of pesticide exposure, showed chlorantraniliprole and cyantraniliprole to be harmless and spinetoram and lambda-cyhalothrin to be harmful to both larvae and adults. While our study with H. convergens differed in being focused on acute mortality from individual exposure routes, it does help to explain the variation in toxicity among natural enemy species and life stages observed from these earlier studies on mortality effects from simultaneous routes of exposure. Chlorantraniliprole and cyantraniliprole were harmless to lacewing larvae and to larval and adult stages of G. occidentalis, but were harmful to lacewing adults and moderately harmful to harmful for larvae and adults of H. convergens. From our study, we found that for H. convergens the most important exposure route for these two insecticides was via oral ingestion, suggesting that natural enemy species or life stages with sucking mouth parts may escape direct oral exposure, whereas those with chewing mouth parts may experience more toxic effects. In contrast, for spinetoram and lambda-cyhalothrin the most important exposure routes for H. convergens were either topical or residual, which helps to explain the more consistent harmful effect of these two insecticides to a broader range of natural enemy species and active life stages. From our study with H. convergens, we have found that even reduced-risk insecticides can, in some cases, cause substantial acute mortality effects in laboratory bio-assays. We have also been able to demonstrate that pesticide, pesticide concentration, predator life stage and exposure route can all influence the acute mortality response of H. convergens in laboratory bio-assays. Of the seven pesticides tested, lambda-cyhalothrin caused the greatest acute mortality for H. convergens, and was classified as either moderately harmful or harmful at the full field rate concentration. Both cyantraniliprole and spinetoram were also either moderately harmful or harmful to H. convergens larvae at the full field rate concentration.