The function of supplemental foods for improved crop establishment of generalist predators Orius insidiosus and Dicyphus hesperus

As with many biological control agents, generalist predators rarely survive prolonged periods of prey scarcity. Towards improving crop establishment of two major predators used in North America, Orius insidiosus and Dicyphus hesperus, this study examined the role of supplemental foods in achieving greater predator survival and faster development. In controlled environment trials, developmental time and survival were compared for predators offered diets including Ephestia eggs, Artemia cysts, Typha pollen, or combinations of these. Nymphal developmental time was significantly shorter and survival greater for both predators reared on diets that included Ephestia eggs. Interestingly, D. hesperus could successfully complete nymphal development on Artemia cysts whereas O. insidiosus could not, alluding to fundamental physiological differences between these predators. In greenhouse assays, D. hesperus was more abundant after six weeks when offered diets that included Ephestia eggs either alone or in combination with pollen or Artemia cysts relative to other diets. In contrast, only diets of Ephestia eggs, Typha pollen or their combination could significantly increase O. insidiosus crop abundance relative to the unfed control. Together, this work highlights important differences in the relative values of supplemental foods for generalist predators used in crop protection. It is also meaningful in guiding biocontrol practitioners globally in the rapidly growing sector of greenhouse vegetable production.

natural colonization of greenhouse crops by mirids Dicyphus tamaninii (Wagner) and Macrolophus caliginosus (Wagner) is regularly observed 18 . In a similar way, the anthocorid predators within the genus Orius (Hemiptera: Anthocoridae), including O. majusculus (Reuter) also naturally colonize and reproduce on greenhouse crops in northern Europe where they attack pests such as thrips 19 . These examples highlight the intrinsic suitability of generalist predatory bugs for biological control of multiple crop pests 18 . In North America the generalist predators Dicyphus hesperus (Knight) and Orius insidiosus (Say) take on a similar function in combatting both established as well as potentially invasive pest species 20 , and it is therefore important to determine how best to optimize their crop presence despite periods of prey scarcity 21 .
For anthocorid and mirid predators, food supplementation with plant or alternate prey resources can greatly improve their long-term crop establishment and contribute to a more sustainable pest management strategy 10,[22][23][24][25] . Traditionally, eggs of the Mediterranean flour moth, Ephestia kuehniella have been widely used as a food source to maintain populations of predatory bugs [26][27][28][29] . However, due to the technical challenge of producing this food source 20 , its cost is generally considered prohibitive for wide scale application in crop food supplementation. As such, the search for less costly yet nutritionally suitable alternatives has led to the investigation of pollen and Artemia cysts as food sources with potential value.
Pollen is also an important food source for a number of flower dwelling anthocorid and some mirid species and its availability on a crop can improve the persistence of these natural enemies 19 . For instance, the fecundity of Orius albidipennis (Reuter) is significantly greater when this predator is offered a combination of pollen and eggs of Ephestia kuehniella relative to Ephestia eggs alone 23 . Orius insidiosus and Macrolophus pygmaeus (Rambur) are also both known to develop to adulthood when reared on pollen alone 10,30 , and O. insidiosus population growth can improve considerably when offered bee collected or Ricinus spp. (Malpighiales: Euphorbiaceae) pollen 29 . However, response to pollen is variable even among members of well recognized pollen feeding predatory species and should be formally evaluated 19 .
Likewise, the cysts of brine shrimp, Artemia spp. (Anostraca: Artemiidae), have also been shown to provide nutritional benefit to some, but not all generalist predators. Orius laevigatus (Fieber) females fed on the cysts of A. fransiscana (Leach), had a similar fecundity relative to such predators offered Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) eggs 31 . In biochemical analyses, Artemia cysts were shown to contain equal or greater amounts of protein relative to eggs of Ephestia, though fatty acid content was nearly three times lower in Artemia cysts relative to Ephestia eggs 24 . These attributes may incur different consequences for predators on a species basis and may affect factors such as the rate of nymphal development, survival, as well as predator fecundity. In laboratory assays, nymphal survival of O. laevigatus was no different between individuals reared on decapsulated Artemia cysts relative to those reared on E. kuehniella eggs 24 . However, developmental time of O. laevigatus was up to 1.6 days longer than individuals reared on E. kuehniella 24 . In this study, laboratory and small cage greenhouse assays were conducted to better understand the relative suitability of Artemia cysts, cattail pollen Typha latifolia L. (Poales: Typhaceae), eggs of Ephestia kuehniella or combinations of these to the nymphal development time, survival and crop establishment of two generalist predators, O. insidiosus and D. hesperus whom are frequently used for biological control of pests on greenhouse crops in North America. The implications for these results are particularly relevant to the development of reliable augmentative biological control strategies involving these species in North America and their economic feasibility is further discussed in light of these results.

Results
Laboratory developmental assays. Dicyphus hesperus. In laboratory trials, the percent survival of developing D. hesperus nymphs varied considerably among predators offered one of the six diet treatments tested, with 83.3% of individuals completing nymphal development on a diet of Ephestia eggs + Artemia cysts, 79.2% on a diet of Ephestia eggs alone and Ephestia + Typha pollen. On an Artemia cyst diet, 66.7% nymphs survived to adulthood compared to only 12.5% on pollen and no survival under control unfed conditions (Fig. 1a). Nymphal developmental time of D. hesperus predators also differed significantly based on supplemental food type (F 4,29 = 4.95, P = 0.0032). Individuals reared on Ephestia eggs alone or combined with either Typha pollen or Artemia cysts developed in the shortest amount of time, followed by those reared on Artemia alone, and last on pollen alone (Table 1).
Orius insidiosus. In a first round of laboratory assays, we evaluated the suitability of various food sources to support the survival and development of O. insidiosus from newly emerged nymphs to adulthood. In general, the survival of O. insidiosus varied widely for each of the different food sources tested. For instance, survival to adulthood was greatest (82.0%) for O. insidiosus reared on eggs of E. kuehniella followed by those reared on a diet containing both Ephestia eggs and pollen (50.0%) or Ephestia eggs and Artemia cysts (42.9%) at a ratio of 1:5 (Fig. 1b). In contrast, nymphs of the unfed control, pollen or Artemia spp. cyst treatments failed to survive past the second, third or fourth instars respectively. As such, comparison of diets for impact on O. insidiosus developmental time from N1 to adult could only be performed for insects successfully reared on three diet treatments: Ephestia eggs, Ephestia eggs + Artemia cysts at a 1:5 ratio, or Typha pollen + Ephestia eggs. Among these three diets, total mean nymphal developmental time for O. insidiosus did not differ significantly (F 2,4 = 0.178, P = 0.8391) and lasted an average of 12.7 ± 0.3 (mean ± SEM) days for individuals reared on Ephestia eggs, 12.3 ± 1.2 days for those reared on Ephestia + Artemia and 12.3 ± 0.5 days for those reared on a diet of Ephestia + Typha pollen (Table 2).
In a second round of laboratory trials, mean total nymphal development time was measured for O. insidiosus reared on one of four different diet treatments relative to an unfed control with each of the four diets offered to predators at five food square rates (½, 1, 2, 4, and 8 squares). In this trial, no nymphs survived to adulthood on diets of dehydrated Artemia cysts, rehydrated Artemia cysts and for predators offered these in any of the amounts tested here, as well as for unfed control predators (Fig. 2). As such only Ephestia eggs and Typha pollen diets, on which predators could develop to adulthood were compared for the impact of food rate on total development    Dicyphus hesperus suppression of greenhouse whitefly. In small cage greenhouse assays, the impact of supplemental foods on D. hesperus crop establishment as well as control of the greenhouse whitefly was assessed. In control cages in which D. hesperus was not added, a clear expansion of greenhouse whitefly populations was evident by week two of the trial, resulting in a significantly greater amount of whitefly in these cages relative to all other treatment cages (P =< 0.0001; Fig. 3a). An increasing trend of whitefly numbers in the no predator control cages also persisted through to the trial end and starkly contrasted with the low level of whitefly consistently observed in all treatment cages containing D. hesperus predators. In this trial, all treatments with the presence of D. hesperus had weekly whitefly counts that did not exceed an average of 104 individuals per cage, a level achieved on week five in the Ephestia and Pollen treatment.
Over the course of this trial, the number of D. hesperus predators per plant increased at markedly different rates depending on the type of food source offered to the predators (Fig. 3b). Notably, predators offered diets including Ephestia eggs either alone (18.0 ± 5.45; P = 0.0014) or in combination with pollen (15.0 ± 3.13; P = 0.0044) or Artemia cysts (11.4 ± 2.92; P = 0.0311) were significantly more abundant relative to the number in no food control cages. Furthermore, a diet of Ephestia eggs alone (P = 0.0023) or in combination with Artemia cysts (P = 0.0481) or pollen (P = 0.0072), also led to significantly greater predator numbers relative to the Artemia cyst diet alone. In contrast, diets such as Artemia cysts alone (5.3 ± 1.86; P = 0.999) and Typha pollen alone (3.3 ± 1.05; P = 0.936) produced a similar number of predators relative to the no food control (2.5 ± 0.91; Fig. 3b). It is notable that, among the three top performing diets, the addition of pollen or Artemia cysts to Ephestia eggs did not improve predator establishment relative to the Ephestia eggs diet alone. However, at points such as in week five of the trial, these two supplemental foods alone may have offered some benefit to predators relative to those in the no food control cages (Fig. 3b).
Orius insidiosus suppression of western flower thrips. In greenhouse trials evaluating the impact of diet on O. insidiosus population buildup and thrips suppression, a diet by time interaction was observed for analyses of both predator and thrips populations (Table 3). Notably, in all O. insidiosus predator containing treatment cages, significant thrips population control was achieved on pepper plants relative to the very high thrips numbers recorded in no predator control cages (P < 0.0001; Fig. 4a). Predator suppression of thrips however, was affected by diet type. This was indicated by the significantly lower total thrips numbers observed in the Ephestia egg (P = 0.0222) as well as in the Ephestia egg plus Typha pollen (P = 0.0162) treatment cages relative to cages in which predators were offered no supplemental food. In addition, though a surge in the number of thrips was apparent in Artemia cyst treatment cages in week five of this trial, no overall trial difference in thrips numbers was observed between this treatment and others except for with the no predator control (P < 0.0001; Fig. 4a).
When examining the numbers of O. insidiosus predators per plant over time among supplemental food treatments, it was clear that diets that included either Ephestia eggs, Typha pollen or combinations of these generally supported greater predator numbers relative to other diets (Fig. 4b). For instance, the highest mean number of O. insidiosus (3.83 ± 0.62 predators per plant over the course of the trial) was recorded in the Ephestia + Typha pollen treatment, equivalent to a trial average of 3.2 times more predators in these cages relative to levels in unfed control cages. This treatment was followed in decreasing rank of predator establishment by treatments of Ephestia eggs alone (3.09 ± 0.53), Typha pollen alone (2.2 ± 0.55), Ephestia eggs and Artemia cysts (1.65 ± 0.5), Artemia alone (1.20 ± 0.43) and finally the no supplemental food control (1.19 ± 0.44). Despite that the addition of pollen to a diet of Ephestia resulted in a numerical increase in predators recorded per plant, no difference (P = 0.433) was observed between these two treatments. Clearly, Typha pollen alone did not have the same nutritional value as Ephestia eggs to O. insidious, as the combination of Ephestia and pollen resulted in significantly greater predator numbers relative to pollen alone (P = 0.0027). However, the addition of pollen did improve to some degree the survival of predators relative to the unfed control (P = 0.0221). Of all food types tested here, Artemia cysts ranked as the least beneficial food source to O. insidiosus with no improvement observed in the number of predators per plant relative to the no food control and no improvement either when this diet was added to another type.

Discussion
The challenge of creating a sustainable augmentative biocontrol program relies on meeting the nutritional requirements for survival and developmental needs of agents applied to crops. This includes mediating the effects of periods of food scarcity and compensating for any nutritional shortcomings of host crops such as the low availability or absence of pollen. Through both highly controlled laboratory assays and more directly translatable greenhouse crop trials, this study served to demonstrate the distinct values of a number of supplemental food types with which it may be possible to achieve greater survival and reproduction of O. insidiosus and D. hesperus predators on commercial greenhouse crops. From this study, it is clear that these two predators respond in very different ways to certain food types and these differences will likely influence how food supplementation is applied in commercial settings.  In laboratory and greenhouse assays the developmental time and percent survival of D. hesperus and O. insidiosus were investigated under different supplemental food regimens. As expected, it was found that diets that included eggs of E. kuehniella invariably sustained the highest survival and shortest nymphal developmental time for both of these predator species. These results are consistent with studies conducted with M. pygmaeus for  which a diet of Ephestia eggs increased the total number of predators on tomato crops relative to control predator population growth 22 as well as for O. insidiosus for which this diet resulted in a greater oviposition rate relative to a diet of pollen or no food at all 29 . Eggs of E. kuehniella are known to have both high protein and high fatty acids contents that together serve to promote predator development and fecundity 24 .
Comparison of alternate food treatments in greenhouse assays with D. hesperus showed that predator establishment on any diet that included Ephestia eggs, even when it was offered as a 1:5 ratio with Artemia cysts offered the greatest level of predator survival. This result, in combination with the relatively comparable nymphal developmental time and survival of predators in laboratory trials involving Artemia cysts alone or Artemia cysts in combination with Ephestia eggs, support the notion that Artemia may also be a suitable food that can help support the development and crop establishment of D. hesperus when it is combined with a small fraction of Ephestia eggs.
This study served to demonstrate the species dependent suitability of much less costly dietary supplements relative to Ephestia eggs. As observed in this study, additives such as Artemia cysts could in some instances contribute to achieving comparable rates of predator development and survival relative to predators offered an Ephestia egg diet. For instance, developmental time of D. hesperus was significantly shorter on Artemia cysts relative to predators reared on pollen or to control unfed predators. Nymphal survival was also good on this food source indicating that it is beneficial to this mirid predator species. In contrast, development of O. insidiosus on Artemia cysts alone did not result in nymphal survival to adulthood and also tended to prolong nymphal development time for the instars predators that did survive. One factor that could have affected this result was that predators were offered dehydrated as opposed to rehydrated Artemia cysts. However in trials performed that examined the value of both dehydrated and rehydrated forms of these cysts to O. insidiosus development and survival, it was found that cyst rehydration did not improve either of these parameters, and no predator was able to survive to adulthood on either diet regardless of the amount. Furthermore, the dehydrated form of Artemia cysts reflects the state in which the food would typically be applied to a commercial greenhouse crop and so remains the most relevant form to test for its nutritional value to predators. For such dehydrated and decapsulated cysts, the average water content is of approximately 3.39% of the total cyst weight. However it is assumed that these cysts would also absorb ambient water molecules over time following application within a given greenhouse crop. In contrast, other studies done to investigate the value of Artemia cysts to Orius thripoborus Hesse and Orius naivashae Poppius have previously been performed with rehydrated cysts, and in those cases it was found that rehydration could facilitate consumption as it provides a moisture source important to the extraction of nutrients by predators 25 . It is therefore likely that the value of Artemia cysts differs considerably for Orius predators on a species basis.
As with other mirids, D. hesperus predators may also have the digestive capacity to allow it to overcome the challenge that a dehydrated Artemia cyst may pose. Dicyphus hesperus is considered an omnivorous predator, that can undergo trophic switching, able to feed both on plant and animal prey sources and using plant moisture to digest prey animals 14,32 . Mirids, Anthocorids and other phytophagous bugs further use their piercing stylets to secrete digestive enzymes into food sources with the goal of disrupting cell membranes, liquefying contents and extracting the most out of their food 33 . Termed extra-oral digestion, this form of nutrient acquisition may be different enough in efficiency in O. insidiosus for digestion of foods so as to render dehydrated Artemia cysts less nutritious 25 . Despite this possibility however, the comparable developmental time of O. insidiosus nymphs when reared on a combined diet of Ephestia and Artemia at a ratio of 1:5 was statistically similar to individuals reared on Ephestia eggs alone, this despite offering predators the same surface area of food in laboratory assays. This indicates that the amount of Ephestia eggs being offered to O. insidiosus per individual could be scaled back through the use of a combined Ephestia egg and Artemia cyst formulation, at considerable cost savings to a biocontrol practitioner.
In terms of the economic feasibility of the various supplemental foods, the integration of Artemia cysts into systems for rearing and improving predator crop establishment has some merit over the sole use of Ephestia eggs. A product currently available in Europe based on cysts of Artemia is currently proposed to be about 3% of the cost of pure Ephestia (Biobest; Beneficial Insectary, Redding, CA). As a compromise between the two food sources, a formulation including one part Ephestia eggs for five parts Artemia cysts (Nutrimac TM Biobest, Westerlo, Belgium) is also available in Europe which is nearly one fifth the cost of pure Ephestia eggs. With this considerably divergent cost in mind, it is more economically feasible to include this alternative food along with Ephestia, particularly as part of a biocontrol program aimed at improving the crop establishment of D. hesperus.
In laboratory and greenhouse trials, Typha pollen alone or in combination with Ephestia eggs did not offer any noticeable nutritional benefit for D. hesperus predators. As a predominantly leaf feeding and dwelling species, it is not surprising that pollen was not considered a relevant food source for this predator. In contrast however, in laboratory and greenhouse trials, this study showed that O. insidiosus survival and development to adulthood could be completed on Typha pollen alone, and that a combined diet of Ephestia and pollen could have positive implications for the establishment of this predator on crop plants. As a natural food source for multiple anthocorid predators, there is inherent logic to the application of both of these foods together, which may prove to be nutritionally complementary for such a flower dwelling and pollen dependent species as is O. insidiosus. In a parallel study, corn pollen was found to increase in O. insidiosus egg production and so possibly other pollen types may also have some value to this predator 34 . However, not all life history parameters of O. insidiosus such as fecundity and longevity are improved by including pollen along with lepidopteran food eggs, so the plant origin of the pollen may be an important factor to consider in future studies 35 .
This study demonstrates the importance of assessing the value of supplemental foods at both the laboratory and greenhouse scales, as each contributes to an understanding of attributes unique to each food source and predator species that may not be obvious otherwise. It was shown that Artemia cysts could be effectively used to promote D. hesperus predator development and that pollen has a role to play in the conservation of O. insidiosus populations, particularly when applied along with Ephestia eggs, or possibly when available as a supplement to pest food resources when available. This knowledge is particularly relevant in the face of growing demand for improved biological control of native crop pest species. It is also relevant in light of a tendency for increased threats by invasive pest species for which generalist predators are frequently seen as a highly adaptable and effective source of biological control [36][37][38][39] . Together, this study plays an important role in demonstrating the function of supplemental foods in improving the resiliency of pest management programs for two important greenhouse crops. They demonstrate the inherent value of such new resources that can be used to counteract pest pressure from unforeseen pest migration events or due to invasive pest introductions. In light of improving the rate of adoption of such supplemental food strategies, future research should be conducted that apply these resources to the commercial greenhouse scale. Trials should also be conducted that focus on identifying the best way to deliver predator foods to the crops without obstructing commercial practices including crop maintenance as well as considerations for pathogen management.

Methods
Insect colonies and plants. Dicyphus hesperus predators were obtained from GrowLiv (Leamington, ON) as adults aged one week or younger. These predators were reared on whole Nicotiana tabacum L. and Verbascum thapsus L. plants kept in 0.5 × 0.5 × 1 m screened cages and maintained in a climate controlled chamber set at 24 °C, 55% RH and a 16 L:8D photoperiod (Conviron A1000, Controlled Environments Ltd. Winnipeg, MB). Dicyphus hesperus predators were fed eggs of E. kuehniella which were applied twice a week directly to plant leaves within the rearing cages. Plants were monitored daily for the emergence of newly hatched (<24 h old) first instar nymphs, which were subsequently, used in laboratory and greenhouse assays.
Orius insidiosus colonies were established from adults supplied by GrowLiv (Leamington, ON). Insects were maintained in a 3 L clear plastic rearing chamber with a mesh-screened lid. Within the rearing chamber, sterile buckwheat hulls served as a niche, and eggs of E. kuehniella were provided onto four 2 cm 2 Post-it ™ note strips (3M, London, ON) with the adhesive surface covered in eggs. Typha pollen was also offered as a food source that was sprinkled directly into the container. These foods were changed twice a week. The colonies were maintained in a climate controlled chamber set at 24 °C, 55% RH and a 16 L: 8D photoperiod. A water wick made from a glass vial filled with moistened cotton as well as beans of Phaseolus vulgaris L. (cv. u1520) served as sources of water, and in the case of the bean, as an oviposition substrate. Fresh beans were provided daily after first being sterilized in a 5% bleach solution for ten minutes. Old beans were removed from the container after a 24 h oviposition period and checked for oviposition marks. Oviposited beans were then maintained in a clear plastic 473 mL cup (Solo ® , Merchants Paper, Windsor, ON,) at 24 °C, 55% RH and a 16 L: 8D. Newly hatched (<24 h old) first instar nymphs were subsequently collected and used in laboratory and greenhouse assays.
Laboratory assays. Impact of diet on predator development time and survival. Laboratory trials were conducted to evaluate the effects of various supplemental foods on predatory bug, D. hesperus and O. insidiosus developmental time and survival. Diet treatments compared in this study included locally-collected fresh-frozen cattail pollen, Typha latifolia, flour moth eggs, E. kuehniella (Beneficial Insectary Inc., Redding, CA) and decapsulated and dehydrated Artemia spp. brine shrimp cysts (Nutrimac TM , Biobest, Westerlo, Belgium). In order to test potential complementarity between diet types, additional treatments included combinations of Artemia cysts with Ephestia eggs at a respective ratio of 5:1 (Nutrimac-Plus TM , Biobest, Westerlo, Belgium) as well as a diet of Ephestia eggs combined with Typha pollen. Each of these five treatments and one unfed control were evaluated for their effects on predator survival and development time through 24 replicated cup assays for trials involving D. hesperus, and two sets of 12 replicated Petri dish assays involving O. insidiosus, the second of which is described below. Dish assays consisted of a fresh pepper leaf disk embedded in cooled but not yet solidified water agar that was aliquoted into a 5 cm diameter Petri dish. The cup assay consisted of a clear plastic 473 mL cup (Solo ® , Merchants Paper, Windsor, ON) which contained a tomato leaflet held in place by a second smaller 59 mL transparent plastic cup (Solo ® , Merchants Paper, Windsor, ON), which was pierced at the bottom to accommodate the leaf petiole. Together these two cups formed a space below that accommodated a 40 mL water reservoir for the leaf and a larger chamber above in which predators could be monitored individually through their nymphal development.
Before adding predators to their respective experimental arena, dry treatment diets were bound to the adherent surfaces of 1 cm 2 or 0.5 cm 2 plastic Post-it ® flags (3M, Maplewood, MN) for trials involving D. hesperus or O.
insidiosus respectively 40 . For each 1 cm 2 flag, the amount of food quantity delivered was measured for a total of ten squares and the average weight (mg ± SEM) is reported as 25.98 ± 1.58 mg for Artemia cysts, 20.92 ± 0.72 mg for Ephestia eggs, 13.74 ± 0.65 mg for the mix of Ephestia eggs and Artemia cysts at a ratio of 1:5 and 3.53 ± 0.16 mg for Typha pollen. With the exception of the Artemia cyst and Ephestia egg combination, combined diets were offered as two separate diet squares at these corresponding weights. Diet squares were replaced at least twice a week to ensure their constant availability and freshness. Insects were transferred individually to experimental arenas, which were then sealed with ventilated plastic lids and maintained in a controlled environment cabinet set at 24 °C, a 16 L: 8D photoperiod and 55% RH. Nymphs of both species were monitored daily for survival, and the timing of moulting events was recorded until the day predators reached adulthood. Statistical analyses were performed to compare the development time of insects maintained on each of the diet treatments. Data were analysed using a one-way ANOVA and means separated using a Tukey's HSD test at α = 0.05 in SAS (v9.3, Cary, NC) 41 .  30 . Also, to investigate the possibility that predator survival was affected by the state of hydration of Artemia cysts or the amount of any of the foods offered to O. insidiosus, trials were conducted that included these variables as factors. Rehydration was accomplished by soaking the cysts in deionized water for two hours. Once again, diet was provided on a Post-it ™ (3M, London, ON) flag cut into 0.5 cm 2 squares. Rates of ½, 1, 2, 4, and 8 squares per arena were tested for each diet type, as well as a control of no food for each diet. For each of the six rates (0, ½, 1, 2, 4, and 8 squares) and four food type treatments, these trials were replicated a total of 12 times. Diet squares were replaced at least twice weekly. In these trials, areas of bare agar within the Petri dish arena were covered with sterile sand so as to minimize the risk of nymphal mortality due to adhesion to sticky agar surfaces (Fig. 5a). Statistical analyses were conducted for data collected from these trials, consisting of comparison of length of complete O. insidiosus nymphal development on the different diet types and amounts through two-way ANOVA and means separation using a Tukey's HSD test at α = 0.05 in SAS (v9.3 Cary, NC) 41 . For these laboratory analyses data were square root transformed to meet the assumptions of analysis of variance. Homogeneity of variance was also confirmed through graphical visualization of residuals against predicted values.

Impact of Artemia cyst rehydration and diet quantity on
Greenhouse trials. In summer and fall 2016, two independent greenhouse trials were conducted, each within two separate greenhouse compartments at the Harrow Research and Development Centre (HRDC) at Agriculture and Agri-Food Canada. Summer trials examined the effects of food supplementation on the establishment of O. insidiosus on pepper crops and on the biological control of this predator on western flower thrips, Frankliniella occidentalis Pergande (Thysanoptera: Thripidae) (Fig. 5b,c). Fall trials focused on assessing the effects of food supplementation on the establishment of D. hesperus on tomato crops and on the biological control of the greenhouse whitefly, Trialeurodes vaporariorum Westwood (Hemiptera: Aleyrodidae leaves so as to prevent food from rolling off plants. For pollen treatments, pollen was directly tapped from the tube, on to the tops of dry leaves (Fig. 5b). Food was replenished onto experimental plants on a weekly basis, directly following plant sampling for predators and pest prey numbers. Impact of diet on Orius insidiosus greenhouse crop establishment. On May 2, 2016, seeds of bell pepper, Capsicum annuum L. (Solanales: Solanaceae) var. Fascinato, were sown onto rockwool plugs, were transferred to cubes after three weeks and were transplanted onto rockwool half-slabs (Grodan, Roxul Inc., Milton, ON), at a rate of two plants per half-slab. One half-slab was then placed within each of the trial cages, six weeks later on June 13 th (Fig. 5). At that time, six of these flowering transplants were randomly sampled to ensure that pre-trial arthropod numbers were low to absent. Briefly, these plants were lightly tapped over a white plastic tray and the number of thrips or other potential insects counted. On June 15 th , adult female western flower thrips, F. occidentalis, were released within each trial cage at a rate of 10 mated females per cage. Thrips were left to establish on plants for a two week period, after which thrips establishment was checked by sampling three cages per greenhouse, once again by lightly tapping plants over a white plastic tray and by counting all thrips from these. Once counted, thrips were placed back onto plants by tapping the tray overtop of the plants. A second release of western flower thrips was repeated on July 26 th at a rate of four mated female thrips per cage. On June 22 nd , 14 O. insidiosus predators were released at a rate of seven fifth instar nymphs and seven adults per cage (three female and four male). On this same date, the supplemental food treatments were distributed onto plants as previously outlined. On June 29 th , one week following the predator release into cages, the first non-destructive sampling for predators as well as for prey thrips was performed. Here also, sampling consisted of lightly tapping plants over a white plastic tray and counting all thrips and predators present. All arthropods were placed back onto plants once counting was completed. Thrips and prey counted included all nymphs and adults of O. insidiosus as well as all nymphs and adult thrips. Plants within these trials with O. insidiosus and thrips on pepper plants were monitored for a total of eight weeks.

Impact of diet on
Statistical analyses of greenhouse trials. The effects of supplemental diet treatment, time and greenhouse block on the number of predators and target pests per cage in each of the pepper and tomato crop trials were compared through repeated measures ANOVA using the PROC MIXED contrast in SAS (v9.3, Cary, NC) 41 . An autoregressive (AR) covariance matrix was used to define the relationship between time points (Table 3). For these analyses, the total number of western flower thrips, F.occidentalis and O. insidiosus predators were analyzed from pepper trials and the total number of greenhouse whitefly, T.vaporariorum and D. hesperus predators were analyzed from tomato trials. Sample units were defined as a cage and the spatial dependence between the cages within each greenhouse was accounted for by including the greenhouse effect as a random block factor in the model. In the case of a significant treatment by time interaction, means separation was performed with Tukey's HSD. For all analyses involved in greenhouse trials, data were square root transformed to meet the assumptions of analysis of variance. Homogeneity of variance was confirmed through graphical visualization of residuals against predicted values.

Data Availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.