Sesame as an Alternative Host Plant to Establish and Retain Predatory Mirids in Open-Field Tomatoes

The silverleaf whitefly (Bemisia tabaci) and the South America tomato pinworm (Tuta absoluta) are two of the most destructive pests of tomato. Open-field tomato production frequently relies on chemical treatments, which has been shown to lead to pesticide resistance. The integration of biological control using predatory mirid bugs is an effective alternative method for managing these pests. However, methods to establish and maintain populations of zoophytophagous mirids are not adequately described. We explored the potential use of two mirids naturally occurring in Florida, Nesidiocoris tenuis and Macrolophus praeclarus. We conducted 6 field experiments over 4 consecutive years to develop a strategy to maintain the mirids. Pre-plant inoculation of tomato plants did not lead to their establishment, likely due to the low prevalence of prey. We explored the use of sesame (Sesamum indicum) to retain the mirids. Intercropping sesame maintained the populations of N. tenuis throughout the duration of the crop. Macrolophus praeclarus never established in any of the open-field experiments. Nesidiocoris tenuis damage was minimal (<1 necrotic ring/plant) and mirid damage was reduced in the presence of sesame. Our results show that intercropping sesame may provide a means to utilize mirids to manage B. tabaci, an established pest, and provide options to tomato growers should T. absoluta invade USA.


Introduction
Tomato, a crop of major economic importance [1], is under constant threat from arthropod pests that can cause severe losses. Among these, the siverleaf whitefly, Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) and the South America tomato pinworm, Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae) are currently two of the most destructive and difficult pests to manage [2,3]. Bemisia tabaci has invaded the United States, Australia, Africa, and several European countries. Bemisia tabaci causes damage through feeding and, more importantly, as a vector of viruses [3]. Tuta absoluta, a pest native to South America, invaded the Mediterranean region in 2006 causing high yield losses [4]. From there, T. absoluta quickly spread across the Afro-Eurasian supercontinent, Central America, and China, and threatens to invade Mexico and the United States, which together account for 25% of the world's tomato crop area and 42% of the world's tomato production [2].
Insecticide application continues to be the primary control method for these pests in most parts of the world [2]. Chemical tools can provide relief, but alternatives are needed because of insecticide resistance and sub-lethal effects [5,6], the ever-increasing close interface with suburban communities, environmental contamination, worker and feeding damage to tomato plants compared to N. tenuis and that adding sesame increased N. tenuis populations without increasing damage to the tomatoes. Although the addition of sesame also slightly increased M. praeclarus numbers in the absence of prey, this species did not reach the adult stage by feeding on sesame alone. In this study, we explored the use of sesame to retain and increase populations of unconfined mirids. Our main goal was to examine the mirids' effectiveness in controlling populations of an established pest, B. tabaci, and to learn how to exploit them for developing an IPM program for the possible incursion of T. absoluta.

Pre-Plant Inoculation of Tomato to Establish N. tenuis in Open Field Tomatoes
Nesidiocoris tenuis dispersed early from the field plots, which corresponded to the decrease in the availability of B. tabaci as prey ( Figure 1C,D). The number of B. tabaci found in plots receiving N. tenuis was similar to both the insecticide-treated and the untreated plots ( Figure 1B) and there was no interaction between week and treatment (F 14,63 = 1.17 p = 0.323). Mirid feeding reduced the subsequent generations of B. tabaci as seen with lower numbers of eggs and nymphs 7 and 8 weeks after release, but their control did not persist ( Figure 1B). Significantly fewer B. tabaci eggs and nymphs were seen on plants treated with insecticides compared to the untreated plots ( Figure 1B, treatment: F 2,9 = 5.85 p < 0.024) as the populations of B. tabaci increased during the study (week: F 7,63 = 12.53 p < 0.001). The mirids did little damage to the tomatoes in accordance with their low numbers. Necrotic rings were observed only on one date (16 March 2015) and on only three plants out of the total 140 plants sampled.
In contrast, N. tenuis was more effective and more damaging when confined to cages (Supplemental Figure S1). The mirids rapidly decimated the number of B. tabaci (treatment: F 2,9 = 53.04 p < 0.001), which remained near zero for the duration of the study (week: F 4,36 = 1.79 p = 0.152, treatment x week F 8,36 = 7.23 p < 0.001). The mirids controlled B. tabaci numbers to similar levels as the insecticide treatment, which was significantly lower compared to cages with B. tabaci-infested tomatoes (S1B, treatment: F 7,24 = 3.13 p = 0.017). Unlike the field study, the number of mirids was consistent throughout the experiment (S1C, F 7,24 = 3.13 p = 0.017). When dispersal was not an option, the number of necrotic rings on the tomatoes remained similar through the course of the experiment (F 3,12 = 0.41 p = 0.748).

Tomato-Sesame Intercrop to Retain Zoophyatophagus Mirids 2.2.1. Field Evaluation of Sesame to Retain N. tenuis in Tomato
The addition of sesame to a tomato field planting retained N. tenuis for the duration of the study even when whitefly numbers were very low ( Figure 2B). The mirids were able to control the initial population of whiteflies as well as natural migration into the field. The mirids reduced the number of whiteflies to near zero levels and the number of whiteflies did not increase during the study ( Figure 2B, F 3,12 = 0.24 p = 0.865). Similar numbers of whiteflies were found on sesame and tomato plants and there were no differences whether the tomato was planted next to sesame or planted next to tomato. Mirids were found on both tomato and sesame plants throughout the study. Mirids were found more frequently in the tomato release plots planted next to tomatoes and in sesame release plots planted next to tomatoes compared to sesame plants planted next to sesame plants and tomato planted next to sesame ( Figure 2C, F 3,12 = 11.36 p < 0.001). The effect of interplanting sesame on tomato damage changed during the study ( Figure 2D, F 9,108 = 18.52 p < 0.001) and there was no interaction between week and treatment (F 9,108 = 1.56 p < 0.06). Initially, more tomato seedlings in the sesame-tomato intercrop had necrotic rings compared to tomato seedlings planted next to tomato. However, the early damage to tomato in the tomato-sesame intercrop was not detectable as the plants grew. In fact, more damaged tomato plants were seen when tomato was planted next to tomato when populations of mirids were high in November compared to the sesame-tomato intercrop that had similarly  Figure 2C), showing the potential benefit of sesame to lower damage to tomatoes as mirid populations increase ( Figure 2D). more tomato seedlings in the sesame-tomato intercrop had necrotic rings compared to tomato seedlings planted next to tomato. However, the early damage to tomato in the tomato-sesame intercrop was not detectable as the plants grew. In fact, more damaged tomato plants were seen when tomato was planted next to tomato when populations of mirids were high in November compared to the sesame-tomato intercrop that had similarly high numbers of plants with mirids ( Figure 2C), showing the potential benefit of sesame to lower damage to tomatoes as mirid populations increase ( Figure 2D).      The spring 2016 field study showed that the amount of sesame needed to retain mirids in the crop could be greatly reduced. With a border of 10 sesame plants, the mirids were retained in the tomato plants even when B. tabaci prey were reduced to near zero ( Figure 3). In fact, the sesame -tomato planting was sufficient to retain the mirids in the tomato-alone planting. There was no N. tenuis damage seen on tomato plants. There were no differences in the number of B. tabaci found in the tomato alone planting and sesame-tomato planting ( Figure 3B, F 1,8 = 4.47 p = 0.066) and there was an interaction between week and treatment (F 6,65 = 12.82 p < 0.001). More whiteflies were found on sesame on the first collection date compared to the tomato plants in the tomato alone or the tomato-sesame intercrop. However, the number of whiteflies was reduced to similar levels as seen on tomato planted alone or in the tomato-sesame intercrop ( Figure 3B). Mirids were seen on sesame and tomato plants 5 weeks after the plants were transplanted into the field and remained at similar levels throughout the remainder of the study ( Figure 3C. week: F 6,54 = 1.76 p = 0.117). There was no difference in the percentage of tomato plants with mirids in the tomato-sesame intercrop or the tomato alone planting and there was no difference in the percentage of sesame and tomato plants with mirids ( Figure 3C, treatment: F 2,9 = 0.533 p = 0.604) and there was no interaction between week and treatment (F 12,54 = 0.413 p = 0.952). The spring 2016 field study showed that the amount of sesame needed to retain mirids in the crop could be greatly reduced. With a border of 10 sesame plants, the mirids were retained in the tomato plants even when B. tabaci prey were reduced to near zero ( Figure 3). In fact, the sesame -tomato planting was sufficient to retain the mirids in the tomato-alone planting. There was no N. tenuis damage seen on tomato plants. There were no differences in the number of B. tabaci found in the tomato alone planting and sesametomato planting ( Figure 3B, F1,8 = 4.47 p = 0.066) and there was an interaction between week and treatment (F6,65 = 12.82 p < 0.001). More whiteflies were found on sesame on the first collection date compared to the tomato plants in the tomato alone or the tomato-sesame intercrop. However, the number of whiteflies was reduced to similar levels as seen on tomato planted alone or in the tomato-sesame intercrop ( Figure 3B). Mirids were seen on sesame and tomato plants 5 weeks after the plants were transplanted into the field and remained at similar levels throughout the remainder of the study ( Figure 3C    The fall 2016 field study showed that N. tenuis released into a tomato-sesame intercrop plots controlled B. tabaci populations on tomato to similar levels found in pesticide-treated plots ( Figure 4B). The experiment also showed the highly mobile nature of the mirids. Nesidiocoris tenuis were found in the untreated tomato plots and insecticide-treated sesame 6 weeks after planting ( Figure 4C,E). This invasion of N. tenuis into the other treatments corresponded to an increase in B. tabaci prey and a subsequent decrease in the number of B. tabaci in untreated plots. Bemisia tabaci numbers increased for 9 weeks and then decreased in the subsequent 4 weeks of the study ( Figure 4B, week: F 12,108 = 25.39 p < 0.001, week x treatment: F 24,108 = 1.43 p = 0.109). There were no significant differences in the number of B. tabaci found on tomatoes in untreated, insecticide-treated and N. tenuis release plots (F 2,9 = 2.40 p = 0.146). The percentage of tomato plants with mirids increased during the study and there were no differences between the amount of tomato plants with mirids in untreated and mirid release plots ( Figure 4C).
The fall 2016 field study showed that N. tenuis released into a tomato-sesame intercrop plots controlled B. tabaci populations on tomato to similar levels found in pesticidetreated plots ( Figure 4B). The experiment also showed the highly mobile nature of the mirids. Nesidiocoris tenuis were found in the untreated tomato plots and insecticide-treated sesame 6 weeks after planting ( Figure 4C,E). This invasion of N. tenuis into the other treatments corresponded to an increase in B. tabaci prey and a subsequent decrease in the number of B. tabaci in untreated plots. Bemisia tabaci numbers increased for 9 weeks and then decreased in the subsequent 4 weeks of the study ( Figure 4B, week: F12,108 = 25.39 p < 0.001, week x treatment: F24,108 = 1.43 p = 0.109). There were no significant differences in the number of B. tabaci found on tomatoes in untreated, insecticide-treated and N. tenuis release plots (F2,9 = 2.40 p = 0.146). The percentage of tomato plants with mirids increased during the study and there were no differences between the amount of tomato plants with mirids in untreated and mirid release plots ( Figure 4C).
Bemisia tabaci and N. tenuis were found on both the untreated and insecticide-treated sesame ( Figure 4D,E). Like the tomato plots, B. tabaci numbers were low throughout the experiment and there were no differences in the number of B. tabaci found in the insecticide-treated and untreated plots (F1,6 = 3.75, p = 0.101). Bemisia tabaci numbers fluctuated during the study with a significant effect of week (F12,72 = 5.64, p =< 0.001) seen, and the lowest number of B. tabaci were found during the last 3 weeks of the study ( Figure 4D). Although the insecticide application did not completely exclude the presence of N. tenuis on sesame, mirid populations were reduced on insecticide-treated sesame with their presence being recorded on only one sample date (28 November 2016). Bemisia tabaci and N. tenuis were found on both the untreated and insecticide-treated sesame ( Figure 4D,E). Like the tomato plots, B. tabaci numbers were low throughout the experiment and there were no differences in the number of B. tabaci found in the insecticidetreated and untreated plots (F 1,6 = 3.75, p = 0.101). Bemisia tabaci numbers fluctuated during the study with a significant effect of week (F 12,72 = 5.64, p =< 0.001) seen, and the lowest number of B. tabaci were found during the last 3 weeks of the study ( Figure 4D). Although the insecticide application did not completely exclude the presence of N. tenuis on sesame, mirid populations were reduced on insecticide-treated sesame with their presence being recorded on only one sample date (28 November 2016).

Field Evaluation of M. praeclarus to Control B. tabaci in a Tomato-Sesame Intercrop
Macrolophus praeclarus were never recovered from tomato alone or the tomato-sesame intercrop. Bemisia tabaci numbers increased in the absence of the mirids. Bemisia tabaci numbers were low for the first 5 sample dates then increased in the subsequent 3 sample dates (week: F 7,84 = 183.8 p < 0.001; week x treatment: F 21,84 = 1.45 p = 0.119).
Although not released, N. tenuis moved into the experimental plots from the surrounding area on the last sampling day. Nesidiocoris tenuis were found in all treatments and there was no effect of planting (tomato alone or tomato-sesame intercrop; F 1,12 = 0.016, p = 0.901) on the percentage of tomato plants with mirids found in the plots ( Figure 5C).  Although not released, N. tenuis moved into the experimental plots from the surrounding area on the last sampling day. Nesidiocoris tenuis were found in all treatments and there was no effect of planting (tomato alone or tomato-sesame intercrop; F1,12 = 0.016, p = 0.901) on the percentage of tomato plants with mirids found in the plots ( Figure 5C). Mirids reduced whiteflies on tomatoes as there were fewer nymphs and eggs in plots with mirids compared to the untreated plots ( Figure 6B, F4,21 = 40.31 p < 0.001). The difference was most evident during the first 5 weeks of the experiment, where the whitefly numbers were reduced by 40-75 % in treatments with mirids compared to untreated tomatoes (week: F7,139 = 41.67 p < 0.001, week x treatment: F28,139 = 2.39 p = 0.001). Plots treated with insecticides had the lowest number of whiteflies, which was consistent through the study ( Figure 6B). Mirids reduced whiteflies on tomatoes as there were fewer nymphs and eggs in plots with mirids compared to the untreated plots ( Figure 6B, F 4,21 = 40.31 p < 0.001). The difference was most evident during the first 5 weeks of the experiment, where the whitefly numbers were reduced by 40-75 % in treatments with mirids compared to untreated tomatoes (week: F 7,139 = 41.67 p < 0.001, week x treatment: F 28,139 = 2.39 p = 0.001). Plots treated with insecticides had the lowest number of whiteflies, which was consistent through the study ( Figure 6B).
Mirid damage was found in all treatments ( Figure 6D). However, the damage was low, with a mean of less than one ring being seen per plant. Corresponding to mirid abundance ( Figure 6C), the number of rings found on tomatoes was highest in treatments with mirid releases (F4,20 = 6.97, p = 0.001). The presence of sesame reduced N. tenuis feeding damage ( Figure 6D). Pesticides reduced but did not totally prevent N. tenuis damage. Nesidiocoris tenuis and necrotic rings were found on pesticide-treated tomato plants on the last two sample dates.

Discussion
The use of zoophytophagous predators in commercial vegetable production has increased worldwide [24]. However, their use in open tomato field cropping systems is very limited compared to greenhouse production. Our results showed that zoophytophagus mirids can be established in open field crops and they can provide control of B. tabaci The highly mobile nature of mirids was very apparent in this experiment. During the first 5 weeks of the study, mirids were present only in plots where mirid releases were made ( Figure 6C). By week 6, N. tenuis was found in all the untreated plots and the M. praeclarus release plots and by week 8 they were present in the pesticide-treated plots. Macrolophus praeclarus were not recovered from the M. praeclarus release plots or from any of the other treatment plots. The establishment of N. tenuis in the untreated treatments paralleled a decrease in Bemisia tabaci numbers ( Figure 6).
Mirid damage was found in all treatments ( Figure 6D). However, the damage was low, with a mean of less than one ring being seen per plant. Corresponding to mirid abundance ( Figure 6C), the number of rings found on tomatoes was highest in treatments with mirid releases (F 4,20 = 6.97, p = 0.001). The presence of sesame reduced N. tenuis feeding damage ( Figure 6D). Pesticides reduced but did not totally prevent N. tenuis damage. Nesidiocoris tenuis and necrotic rings were found on pesticide-treated tomato plants on the last two sample dates.

Discussion
The use of zoophytophagous predators in commercial vegetable production has increased worldwide [24]. However, their use in open tomato field cropping systems is very limited compared to greenhouse production. Our results showed that zoophytophagus mirids can be established in open field crops and they can provide control of B. tabaci when other management strategies were also in place. These results lead to the possibility of developing an IPM strategy utilizing mirids as biological control agents in open-field tomatoes.
Inoculating tomato seedlings with mirids prior to field planting was critical to establish the predator in the field. In our initial field studies, we never recovered N. tenuis when adults were released directly into plots. With low prey densities, the mirids quickly dispersed and providing supplementary food to establish the mirids on tomato seedlings prior to planting was found to be necessary.
Utilizing mirids' phytophagous behavior appears to be a viable means to retain them when prey numbers are low. With the addition of sesame as a companion crop, N. tenuis was present the entire cropping season even when B. tabaci populations were near zero. Nesidiocoris tenuis can reproduce on sesame in the absence of prey [40]. In cage studies, the combination of sesame and tomato increased N. tenuis populations without increasing damage to the tomatoes [17,28,31]. This study showed that the benefits previously seen with T. absoluta when mirids, N. tenuis and Macrolophus pygmaeus, were confined to a cage [31] could also work to control B. tabaci in open-field production where the mirids could leave the crop. When given the ability to move freely in the field, N. tenuis did not remain on sesame despite the presence of whiteflies and being a better plant resource than tomato [27]. The mirid was found equally on tomato and sesame plants and readily moved into to treatment plots containing only tomato. Mirids migrating into tomato crops have been found to concentrate on plants with high prey density, which may enhance pest control on tomato. In all four years, N. tenuis reduced the number of B. tabaci to very low numbers (<10 whiteflies/leaf), which were similar to numbers seen in the pots where pesticides commonly used by USA growers were applied. In fact, N. tenuis began to invade all treatments including the pesticide-treated pots once whitefly numbers began to increase. This indicates that N. tenuis may work effectively to control whiteflies in open fields. However, further work is needed to evaluate whether the mirids can provide control similar to pesticides in large fields, typical of USA, Mexico and China production. Our reduced sesame planting study showed that a number of sesame plants needed to retain the mirid could be decreased and still retain mirids in the plots. Additional studies are needed to determine the amount of sesame and planting distribution to have the mirids remain and effectively control B. tabaci in large commercial tomato fields.
Over the course of this study, N. tenuis was found to establish near the experimental plots. In fact, the mirid invaded the M. praeclarus experiment despite not being released that season. Where N. tenuis is established, the mirid can fortuitously colonize crops [41][42][43]. This is advantageous in Southern Europe as N. tenuis controls T. absoluta [14]. However, their phytophagy can cause significant damage to tomatoes grown in heated greenhouses when pest populations are low [25,44]. Consequently, N. tenuis is considered a pest of heated greenhouse tomatoes in Northern Europe [45]. Nesidiocoris tenuis can also cause damage in open-field sesame and tobacco crops and is considered a pest [46,47]. Our cage studies and many other greenhouse studies showed that confining N. tenuis forces the mirid to rely on tomato plants for survival. Logically, the likelihood for damage increases. Our studies suggest that N. tenuis will disperse and cause minimal damage in open-field tomato cropping systems. In fact, we recorded very low levels of damage and had difficulty maintaining a population in the crop without the presence of sesame. Hence, the status of N. tenuis as pest or biological control agent will depend on the crop, pest complex, and cropping system.
Other species of mirids that cause less damage to tomato have been used to control pests in greenhouse production systems [13]. For the USA, M. praeclarus appeared to be a good biological control candidate. In laboratory and field cage studies, the mirid controlled B. tabaci, preyed on moth eggs (Ephestia kuehniella), and caused no damage when confined to tomato plants [17,32]. In addition, M. praeclarus is reported from South Florida, the Caribbean, Mexico, and Central America, making this species a potentially accessible biological control agent where T. absoluta has not invaded. However, differ-ent methods are needed to utilize M. praeclarus in open-field tomatoes. In our studies, M. praeclarus did not establish in tomato alone or in a sesame-tomato intercrop. On the contrary, N. tenuis invaded the M. praeclarus release plots in both experiments where M. praeclarus was released. Further studies are needed to explore whether M. praeclarus was not recovered because N. tenuis fed upon or behaviorally displaced M. praeclarus or whether other factors in the experiment (i.e., low density of prey or environmental conditions) caused the mirid to disperse.
Pesticide resistance and environmental problems necessitate looking for alternatives to control important tomato pests like B. tabaci and T. absoluta. For successful biological control, predators must establish when the pest population is low [12,21]. As seen in our studies, B. tabaci invasions are unpredictable. The mirids' generalist predation behavior has the benefit of controlling other pests, like tomato leafminers, should they invade the crop. Although N. tenuis and Macrolophus species are generalists, they show prey preference [47][48][49][50]. In fact, the need to provide N. tenuis supplementary food in order to retain the mirids to control T. absoluta is reduced when B tabaci are present [51]. Understanding the mirids' prey preference is essential to select species for release that will likely attack the target pest(s) when given a choice. In addition, the mirid's phytophagous feeding behavior can help to maintain the predator in the field in the absence of the target pest. We showed that a low density of sesame plants served as a method to retain the predator when prey densities were low. Additionally, non-cultivated host plants may provide refuges from pesticide residues [52,53]. Our studies were the first steps in learning how to incorporate zoophytophagus mirids to manage an existing pest, whiteflies, in open field tomatoes as well as prepare for potential new invaders such as T. absoluta that are spreading in the region.

Field Plot Preparation
Six field experiments were conducted to determine whether M. praclearus and N. tenuis could be established in open field tomato production systems and whether the mirids damaged the tomatoes. All trials were conducted at the IFAS /SWFREC research facility near Immokalee, Florida, on single row raised beds 81 cm × 128 m long and 20 cm high. The experiments were run during the south Florida tomato growing season; in the fall (October-January) and in the spring (March-May). Tomato varieties were selected to best match the planting season [54]. Granular 10-2-10 NPK fertilizer was incorporated before each planting at a rate of 121 kg N/ha. The remaining fertilizer requirements were met through fertigation over the course of the crop with liquid 7-2-7 NPK. Beds were then fumigated with 121 kg/ha 50:50 methyl bromide + chloropicrin, two drip tapes with 20.3 cm emitter spacing were laid down and beds were covered with whiteface (fall) or black (spring) polyethylene film mulch (fumigant barrier film, Raven VaporSafe ® 1.1 mil, Raven Engineered Film, Sioux Falls, SD, USA). The experimental rows of plants were separated by two rows (81 cm × 128 m) of corn (Zea mays L.). Experimental plots were arranged in a randomized complete block design (RCB).

Pre-Plant Inoculation of Tomato to Establish N. tenuis in Open-Field Tomatoes
We conducted studies to determine whether N. tenuis would establish in open-field tomatoes by inoculating seedling tomatoes with the mirid and prey (B. tabaci). Tomato plants were inoculated with whiteflies and N. tenuis by placing tomato seedlings (TYLCV tolerant variety 'BHN 8845 ) into cages. To inoculate the tomatoes with B. tabaci, infested sesame plants (5-6) taken from the colony were placed in the cage to allow an estimated 10-15 adults per seedling to move to the tomato plants. For mirid treatment, 50 female and 50 male N. tenuis were randomly collected from the colony and introduced to a second cage with 100 B. tabaci-infested tomatoes as previously described. The mirids were provisioned with E. kuehniella eggs/Artemia cysts and allowed to oviposit for 1 week. The seedlings were transplanted on 2 March 2015, into two rows of raised beds (46 cm spacing/35 plants per plot, Figure 1A), with each tomato plot separated by a plot of corn (10 plants/plot). Seedlings in the insecticide treatment plots received 1 application of imidacloprid (Admire ® Pro 207 mL/4047 m 2 , Bayer Crop Science LP, Research Triangle Park, NC, USA) applied in transplanting/setting water 7 days after transplanting. Whiteflies were also allowed to infest the experimental plots naturally. The experiment consisted of three treatments (n = 4): (1) N. tenuis + B. tabaci, (2) insecticide + B. tabaci and (3) untreated tomatoes + B. tabaci. Whitefly populations were monitored weekly by randomly selecting 5 plants from the middle of each experimental plot. One leaf was randomly selected from the top 25-50 cm of the plant. In the laboratory, all whitefly eggs and nymphs were counted using a dissecting microscope (10×). Mirids were monitored weekly by counting all mirid adults and nymphs in situ seen on each tomato plant in the N. tenuis + B. tabaci plots.

Cage Evaluation of Pesticide and N. tenuis to Control B. tabaci
To understand the impact of the mirids under confined conditions, the comparison of N. tenuis to pesticides experiment was repeated in cages (61 cm × 61 cm × 91 cm, Bugdorm MegaView Science Company, Ltd., Taichung Taiwan) held in a polycarbonate greenhouse provided with evaporative cooling (25 ± 1 • C, 60 % humidity and 14:10 h (L:D) photoperiodic conditions). This experiment used the same pool of whitefly and whitefly+ mirid-infested seedlings as were used for the field experiments. Four potted plants, taken from the same pool of plants used in the field study described above, were randomly assigned to cages, which were then arranged in a randomly completed block design (n = 4) in the greenhouse. As above, the experiment consisted of three treatments: (1) N. tenuis + B. tabaci, (2) insecticide + B. tabaci and (3) untreated tomatoes + B. tabaci. Bemisia tabaci populations were monitored weekly by taking 1 leaf from the middle of each plant. The leaves were taken to the laboratory and examined under the dissecting microscope (10×) and all whitefly eggs and nymphs were counted. N. tenuis populations and associated damage were monitored by randomly selecting one of the plants in the cage and counting all mirids (adults and nymphs) and necrotic rings in situ. In the fall of 2015, a field experiment was conducted to determine whether sesame plants would help retain N. tenuis when planted near tomatoes. Tomato (var: BHN 8846) and sesame seedlings were pre-infested with whiteflies and mirids prior to transplanting in the fields. To infest seedlings with whiteflies, plants were confined in cages with highly infested sesame plants for 1 week. To inoculate the plants with mirids, 100 whitefly-infested tomato and sesame seedlings were placed in cages with 50 female and 50 male N. tenuis (1 mirid/plant) for 1 week. Additionally, the mirids were provided with E. kuehniella eggs and Artemia spp. cysts ad libitum. Each whole plot contained a total of 36 plants spaced 30.5 cm apart (Figure 2A). The center 12 plants were infested with N. tenuis (tomato or sesame) and whiteflies. On either side of the center planting were 12 whitefly-infested sesame or tomato (no N. tenuis) seedlings. There were 4 treatments arranged in a RCB (n = 4): (1) tomato, tomato + N. tenuis, tomato; (2) sesame, sesame + N. tenuis, sesame; (3) tomato, sesame + N. tenuis, tomato; (4) sesame, tomato + N. tenuis, sesame. Beginning 1 month after transplanting, the number of whitefly nymphs and eggs was evaluated each week by selecting 1 leaf from the mid-canopy of 3 randomly selected plants in the central planting. The presence and absence of N. tenuis were evaluated by counting the number of plants in the center sub-plot with at least 1 mirid per plant. N. tenuis damage was assessed by counting the number of tomato plants from the center planting (treatments 1 and 3) with visible necrotic rings.

Field Evaluation of Reduced Amount of Sesame in Tomato-Sesame Interplanting
In the spring of 2016, the amount of sesame interplanted with tomato was reduced to determine whether the lower amount would still help retain N. tenuis. Tomato (var: BHN 8845) and sesame seedlings were infested with B. tabaci by placing trays in cages with highly B. tabaci-infested collards for 1 week. Additional seedlings were infested with B. tabaci and N. tenuis by introducing adult mirids (1 adult/plant) along with the infested collard plants. As above, the mirids were provided with E. kuehniella eggs and Artemia spp. cysts to ensure survival and maximum oviposition. The infested seedlings were transplanted on 4 March 2016, with a 30.5 cm spacing between plants. There were two treatments: tomato alone (40 plants per plot) and sesame-tomato interplanting (10 sesame plants next to 20 tomato plants per plot; Figure 3A). Treatments were separated by 14 tomato plants sprayed on 19 April 2016 with Dinotefuran (Venom 118 mL/4047 m 2 ). The treatments were arranged in RCB (n = 4) distributed in two rows separated by rows of corn. Whiteflies were monitored weekly by taking 1 leaflet from 50% of the plants in the plot. In the laboratory, all whitefly eggs and nymphs were counted using a dissecting microscope (10×). The presence and absence of N. tenuis was evaluated by counting the number of plants in the plot having at least 1 adult or nymph. In the fall of 2016, the ability of N. tenuis to control B. tabaci was compared to insecticidetreated tomatoes in plantings bordered by sesame. Each treatment plot had 30 grape tomatoes ('BHN784 ) bordered by 10 sesame plants and was assigned to 1 of 3 treatments (n = 4): (1) untreated tomato, (2) insecticide-treated tomato or (3) tomato + N. tenuis. Seedlings were transplanted on 19 September 2016, after being inoculated with B. tabaci and N. tenuis for mirid treatment plots 1 week prior, as described above ( Figure 4A). Initial field samples indicated that populations of B. tabaci were low, and additional releases were made by placing 1 B. tabaci infested tomato plant (var. 'Lanai') per plot on 24 October 2016. Mirid numbers were also low and an additional release of 1 N. tenuis per plant was performed on October 31 in the designated mirid treatment plots. The insecticide-treated tomato plots and the sesame bordering the untreated tomato plots were drenched with imidacloprid (Admire ® Pro 207 mL/ 4047 m 2 ). The sesame was treated with pesticides to form a barrier to help limit the possible spread of N. tenuis to the untreated tomato plots. Acibenzolar-Smethyl (Actigard ®  This field experiment evaluated the ability of M. praeclarus to establish and persist on tomato and whether intercropped sesame functioned to maintain M. praeclarus in the field. Two weeks prior to planting in the field, grape tomato (Jolly Girl) and sesame seedlings were infested with whiteflies by placing highly infested collard leaves taken from the colony into cages with the seedlings. The following week, M. praeclarus adults (1 per plant) were released into the cages and allowed to oviposit. As above, the mirids were provisioned with E. kuehniella eggs and Artemia spp. In the spring of 2018, a field study was conducted to study the ability of M. praeclarus and N. tenuis to control B. tabaci in tomato. The previous field studies showed that intercropping sesame with tomato helped retain N. tenuis, but the presence of sesame did not help to retain M. praeclarus. Therefore, tomato-sesame intercrop treatment plots were established with only N. tenuis. Each treatment plot held 30 plants (60 cm spacing) bordered by 10 corn plants ( Figure 6A). There were 5 treatments (n = 5): (1) untreated tomato (grape tomatoes, var. Jolly Girl), (2) insecticide-treated tomato, (3) tomato alone + N. tenuis, (4) tomato alone + M. praeclarus or (5) tomato-sesame intercrop + N. tenuis (5 tomato plants:1 sesame plant). To establish the whiteflies in the treatment plots, tomato and sesame seedlings were placed inside 24 × 61 × 91 cm cages two weeks prior to planting in the field and infested with whiteflies by adding collard leaves taken from the colony. One week later, 1 mirid per plant was released into cages with plants designated for treatments with mirids. The seedlings were transplanted into the field on 13 March 2018. Two weeks after field planting, Verimark (90 mL/plant) was applied to the pesticide treatment plots. To suppress fungal and bacterial disease, extract of Reynoutria sachalinensis was applied to all plots on 27 March, 9 April, and 16 April 2018 at a rate of 0.95 L per 76, 151, and 227 L of water per 4047 m 2 per date, respectively. The fungicide Actigard was applied to all treatments on 27 March, 9 April, and 16 April 2018 at a rate of 10 mL per 76, 151, and 227 L of water per 4047 m 2 per date, respectively. Each week, B. tabaci populations were monitored by sampling 1 leaflet from the mid-canopy of 6 randomly selected plants per plot. In the laboratory, all eggs and nymphs were counted using a dissecting microscope (10×). The presence of mirids was determined weekly by counting the number of plants with at least one mirid adult or nymph per plot. At the end of the experiment, mirid feeding damage was assessed by counting the number of necrotic rings found on 5 tomato stems taken from the top one-third of the canopy of 5 randomly selected plants.

Statistical Analysis
Bemisia tabaci, mirid, and necrotic rings (mirid damage) data were tested for normality (Shapiro-Wilk test) and homoscedasticity (Bartlett test) using software from SAS Institute (JMP, Cary, NC, USA). The total number of egg and nymph B. tabaci and total number of adult and nymph mirids were transformed using log (x + 1) for count data and arcsinesquare root for percentage data to meet the assumptions of analysis of variance (ANOVA). The data were analyzed using a two-factor nested random effects model with treatment and sampling week considered as fixed factors and plot as a random factor. Treatment was nested within plot. Pairwise comparisons of the fixed factor levels were made using the Tukey's HSD post hoc tests. p-values ≤ 0.05 were considered statistically significant.