Field Preparation

There were 3 field tests with 1 test per season in the spring, summer, and fall 2010. All field data were collected in 2 fields; there was one field per location: Kendall and Homestead. Test fields were on Krome gravelly loam soil (loamyskeletal, carbonatic, hypothermic, lithic, udorthents) that was well drained, had a pH of 7.4–8.4, was 34–76% limestone pebbles (≥ 2 mm diam), and had low organic matter (< 2%) (Nobel et al. 1996; Li 2001). Size of each test field was 0.4 ha divided into 40 equal 0.01-ha plots (replications) that were each 2 rows wide by 27.5 m long. Zea mays L. ‘Obsession’ sweet corn (Seminis Vegetable Seeds, Oxnard, California) was planted at Homestead using a garden seeder (Model 1001-B, Earthway® Products, Bristol, Indiana) on 15 Feb, 2 June, and 1 Oct for the spring, summer, and fall tests, respectively. At Kendall, Z. mays ‘Pioneer 3394’ field corn (Pioneer Seeds, Johnston, Iowa) was planted instead of sweet corn. Planting sites were spaced 0.3 m apart in rows separated by 0.9 m. Three to 5 seeds were planted at each planting site to assure that at least 1 seed germinated, but seedlings were thinned to 1 plant per site upon emergence. In parallel bands spaced 0.1 m from the seed rows, granular fertilizer (N-P-K: 8-16-16, Diamond R fertilizers, Fort Pierce, Florida) was applied at 1,347 kg/ha at planting. In addition, a liquid foliar fertilizer spray (N-P-K: 4-0-8, Diamond R fertilizers, Fort Pierce, Florida) was applied twice at 21 and 35 days after planting (DAP) to provide 2.8 kg of N/ ha/day. To control weeds, 0.09 kg/ha of the preemergent herbicide trifluralin (Treflan^TM 4EC, Dow AgroSciences, Indianapolis, Indiana) was applied at planting, and plants were irrigated as needed by drip irrigation. Drip irrigation by a T-Tape system (John Deere Co., San Marcos, California) was provided at 0.704 kg/cm² by 2 parallel lines per row; each line had one hole (2 mm diam) per 15 cm of irrigation line. The system ran for 1 hr/day dispensing 151 L of well water per 30.5 m of row length.

Data Collection

Sweet corn fields were sampled once in each stage of reproductive development (R1, R2, and R3) (Bean 2010) per crop in each season (spring, summer, and fall). The sample in R1 was 4 days after first silk, the second sample was 5 days after the beginning of R2, and the final sample was in the first week of R3. In the spring and fall, samples in the R1, R2, and R3 stages were 49, 63, and 70 DAP, respectively, and in the summer, their respective times were 56, 70, and 77 DAP. For each sample date, 1 corn ear was collected from each of 2 randomly sampled corn plants per plot resulting in 80 ears total. Each ear was placed separately in a 17 × 22-cm, self-sealing plastic bag, and samples were stored at 26 ± 5 °C if they could not be immediately processed. Each ear was then cut in half and placed in a 100-mL beaker of 75% ethanol for 5 min to collect arthropods, then removed with forceps and discarded. Smaller pieces of corn cob, husk, and silk were also rinsed in alcohol to remove arthropods. Alcohol rinses were filtered with a 25-µm mesh (USA Standard Testing Sieve, W. S. Tyler Co., Mentor, Ohio) to collect arthropods, which were saved in 70% ethanol for identification. For field corn at Kendall, only the top 5 cm of ear sheath (including leaf tips and silks but not the cob) were collected and handled as described for sweet corn. A 10X microscope was the instrument of choice for counting arthropods, which were identified to species where possible. Unidentified arthropods were sent for identification and voucher specimens submitted to the Florida Department of Agriculture and Consumer Services, Division of Plant Industry, Gainesville.

Spatial Distribution of Predators and Corn-Infesting Ulidiids

Kalsi (2011) and Kalsi et al. (2014) provided results of the initial abundance and diversity survey. However, the present study also used the results to determine within-field distributions of ulidiid eggs, larvae, and 2 key predators: adults and nymphs of O. insidiosus and larvae of Anotylus insignitus (Gravenhorst) (Coleoptera: Staphylinidae: Oxytelinae) (Fig. 1). Orius insidiosus nymphs, adults, and ulidiid eggs and larvae in corn plants all originated from natural populations at the site where they were surveyed (Kendall or Homestead). However, A. insignitus larvae originated only at Homestead. We applied Taylor's Power Law (b) (Taylor 1961), Iwao's Patchiness Regression (β) (Iwao 1968), and the Index of Dispersion (ID) (Clapham 1936; Selby 1965; Perry & Mead 1979) to quantify aggregation patterns. ID, b, and β are indices of dispersion (aggregation) for a species (Southwood 1978), and each indicates if the distribution pattern is aggregated, random, or regular (uniform), which occurs when ID, b, or β are either > 1, ∼ 1, or < 1, respectively. Using the sample mean () and sample variance (s ²), index of dispersion (ID) values were determined (equation 1). Taylor's power law (equation 2) and Iwao's patchiness regression parameters (equation 3) were calculated using general linear regression models (Southwood 1978; SAS Institute 2003). Taylor power law determines relationships between log , log s², and log a (sampling factor) (equation 2). Iwao patchiness regression relates Lloyd (1967) mean crowding index [(s²/) -1], the sample mean (), and the index of contagion or tendency toward crowding (α) in equation (3). To determine within-field distributions for the predators and ulidiids using the Taylor (b) and Iwao (β) indices, we first determined the goodness of fit to each linear model using regression coefficients (r ²). Then, a student's t-test helped to determine if b and β were significantly different from 1.0: if r ² approximately equals 1, the model provided the best fit.

Fig. 1.

Euxesta stigmatias and 2 of its predators. (A) Eggs of E. stigmatias in corn silk. (B) Third instar larva of E. stigmatias. (C) Adult of O. insidiosus feeding on a third instar larva of E. stigmatias. (D) Anotylus insignitus larva feeding on E. stigmatias eggs.

Functional Response of O. insidiosus to Eggs of Euxesta Species

Ulidiid Colonies to Supply Predators. Each laboratory colony was established with 100 adults in the summer of 2010 with separate colonies for E. stigmatias, E. eluta, and E. annonae collected from corn fields. Each Euxesta sp. was reared in a separate 31 × 31 × 31-cm cage maintained at 27 ± 5 °C and 75 ± 10% RH with rearing methods the same for each fly species. Adult flies were supplied with 1% honey solution and fresh water, and colonies were fed a beet armyworm (Spodoptera exigua (Hübner); Noctuidae) artificial diet (BAW, Southland Co., Lake Village, Arkansas) using the methods of Seal & Jansson (1989). Diet mixture was attached to the bottoms of plastic 28-g vials (BioServe^TM, Beltsville, Maryland, USA) that were placed upside-down on the cage ceilings with the tops facing down to allow oviposition. Eggs were collected at 24-h intervals and then transferred to fresh BAW diet for larval emergence in the same environmental conditions as adults to develop a uniform colony of each fly species. Newly eclosed first-instars were removed from egg containers every 24 h, transferred to different 28-g vials, maintained on BAW diet, and allowed to pupate. Every 4 h, diet vials were checked to remove pupae, which were washed gently with tap water to remove dietary residue and to avoid fungal infection. Pupae were then air-dried and placed into Petri dishes each with a moist, disk-shaped, 5-cm-diam paper to prevent desiccation. Petri dishes with pupae were placed into 31 × 31 × 31-cm cages and checked every 2 h to collect newly emerged adults.

Functional responses were calculated for O. insidiosus to eggs of E. stigmatias, E. eluta, and E. annonae in the laboratory. The tests were conducted in a growth chamber at 27 ± 3.5 °C and 75 ± 10% humidity. Eggs of Euxesta spp. were obtained from the laboratory colonies and O. insidiosus adults were obtained from Koppert Biological Systems (Romulus, Michigan). Each experimental unit was the same kind of 28-g plastic vial used to maintain Euxesta spp. colonies, but the vials remained empty. However, a 3.5-cm-diam parafilm layer was placed over and depressed 2 cm into each vial and provided with Euxesta spp. eggs at specific densities, hence, an area was formed where predator and prey were introduced. Prey Euxesta spp. eggs were killed by cooling for 30 min at 2 °C, then 1-day-old eggs were added to the vials at 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55,and 60 eggs per vial with 1 Euxesta sp. per vial. One O. insidiosus adult was then added to each vial after being starved for 24 h. Two or 3 water drops were added to the paraffin layer of each vial as a moisture source, then the vial was covered with a Petri dish lid. Eggs preyed upon by O. insidiosus became deflated with an air vacuole that did not appear in control vials without the predators, thus, deflated eggs were assumed to have been attacked by O. insidiosus. After 24 h, the number of deflated eggs in each arena was counted, and if the predator died or larvae emerged from eggs, the container was excluded from the test and discarded. Each treatment with a given number of eggs per arena was replicated 8 times.

The predation data allowed us to determine the shape of the resulting curves and the type of functional responses (Type I, II or III), handling times, and attack constants. First, the type and shape of each functional response curve was determined with the polynomial regression model, equation (4) (PROC CATMOD, SAS Institute 2003) in which N _e is the number of prey eaten, N ₀ is the initial number of prey, and P ₀, P ₁ P ₂, and P ₃ are the estimated parameters. For each initial prey density and proportion of eggs eaten, a polynomial equation was determined. If the first parameter (P ₁) was negative, the functional response was type II, and if positive, it was a type III. Once the type of functional response was determined, data were fit to the mechanistic model and the random predator equation, and handling times and attack constants were estimated (PROC NLIN, SAS Institute 2003). A non-linear, least-square regression of the number of flies offered versus number eaten was used to estimate and compare the functional response parameters. The Hassell equation (5) was used to estimate the search rate and attack constant parameters using a constant (b), handling time (T _h), time taken by a predator to search for its prey (a), and the total time available for O. insidiosus to search for and attack its prey (T).

Table 1.

Dispersion indices for ulidiid eggs and larvae, nymphs and adults of orius insidiosus, and larvae of anotylus insignitus in sweet corn ears at silking (R1), blister (R2), and milk (R3) stages in 2010¹.

Spatial Distribution of Predators and Ulidiids

Spring 2010. Based on the Taylor (b), Iwao (β), and Index of Dispersion (ID) variables, distributions of ulidiid eggs in all 3 corn reproductive development stages and ulidiid larvae in the R2 and R3 stages were aggregated (b, β, & ID < 1; P ≤ 0.05; Table 1). Numbers of ulidiid larvae in the Rl stage were not sufficiently large to estimate distribution patterns. Similarly, O. insidiosus had aggregated distributions with b, β, and ID all significantly > 1 during R1 and R2 (Table 1). Also, for each stage and dispersion index, the coefficient of determination (r²) was near 1 indicating a good fit. However, in the R3 stage, O. insidiosus was distributed randomly (ID ∼ 1) or regularly (b and β 1) (Table 1).

Summer 2010. Each dispersion index for ulidiid eggs in the R1, R2, and R3 stages yielded an aggregated distribution (b, β, & ID > 1) (Table 1). During the R1 stage of corn development, numbers of ulidiid larvae were too sparse for distributions to be determined, but in R2 and R3, they had aggregated distributions based on all 3 dispersion indices. The indices b, β, and ID also showed aggregated distributions for O. insidiosus during the R1, R2, and R3 stages (Table 1). Larvae of A. insignitus were not present in sufficient numbers in R1 or R2 to determine dispersion indices, but in R3, they had aggregated distributions based on all 3 indices.

Fall 2010. The indices all yielded aggregated distributions (b, β, & ID > 1) for ulidiid eggs during the R1 and R2 stages except for β in R1, which was random (β ∼ 1). However, there were too few eggs to determine distributions in R3 (Table 1). Ulidiid larvae were too sparse in R1 to determine distributions, but in R2 and R3, larval numbers were sufficient to reveal aggregated distributions based on all 3 indices. All 3 indices showed that O. insidiosus distributions were aggregated during R1 and R2, but they were random (6 & β ∼ 1) or regular (ID < 1) during R3.

Functional Response of O. insidiosus to Eggs of Euxesta spp.

Linear regression equations (Hassell 1978) were calculated for the proportions of eggs consumed per initial number of eggs of E. stigmatias, E. eluta, E. annonae, and their cumulative means (Fig. 2). The first terms of the equations were positive: hence, O. insidiosus yielded Type III functional responses to different egg densities of Euxesta spp. The Hassell equation was therefore used to estimate the search rate and attack constant parameters. Also, the number of prey eggs declined as they were consumed because they were not replaced with fresh eggs, thus, Juliano's (1993) method was used to fit the data to the Rogers (1972) random predator equation. At the greatest prey density (60 eggs per arena), O. insidiosus consumed an average of about 40 Euxesta spp. eggs. For each initial number of eggs offered, there was no significant difference in predation rate for O. insidious among eggs of the 3 Euxesta spp., which also yielded similar handling times (0.43–0.45 h) and attack constants (0.03– 0.05) (Table 2).

Fig. 2.

Type III functional responses of Orius insidiosus to eggs of Euxesta stigmatias, E. eluta, E. annonae, and the combination (mean) of these prey species. Total exposure time was 24 h.

Discussion

Insect populations have 3 types of distributions: random, aggregated, and uniform (Taylor 1961, Southwood 1978). According to Southwood ( 1978), an insect's distribution pattern is greatly influenced by its within-field density. A low capture rate during sampling will often lower the density and indices of dispersal, such as from aggregated distribution to random. However, if the arthropods are caught more often during sampling, density may be higher and hence aggregated instead of random or uniform. Taylor Power Law, Iwao Patchiness Regression, and the Index of Dispersion helped to clarify distribution patterns of ulidiid eggs and larvae and of their predators found on corn ears. Except for the present study and the investigation of ulidiids and Z. longipes adults by Kalsi (2011), these techniques have not been used previously to find spatial distributions of corn-infesting ulidiids or their natural enemies. However, Goyal (2010) used Morisita's index to study distribution patterns of adult E. stigmatias, E. eluta, and C. massyla caught in sticky traps in corn fields from the first appearance of corn silk until harvest. He found that the flies had aggregated distributions on most sampling dates from the first appearance of silk until harvest in corn fields of 1.8 ha and 4–16 ha. In the present study, the distribution pattern of ulidiid larvae was not detected in R1 of any season because they were absent or had low populations. However, similar to Goyal (2010) for adult ulidiids, we found that distributions of ulidiid eggs and larvae on the other sample dates were mostly aggregated.

Table 2.

Handling times and attack constants of Orius insidiosus with respect to eggs of Euxesta spp.¹

The abundance of O. insidiosus increased during summer 2010, when abundances of ulidiid eggs and larvae were also greatest (Kalsi et al. 2014). Other than ulidiid eggs and/or larvae, O. insidiosus was among the most abundant arthropod species during R2 of the summer and fall and statistically the most abundant within R1 of the spring, summer and fall (Kalsi et al. 2014). However, O. insidiosus populations were numerically lowest during R3 of each season (Kalsi et al. 2014). Knutson & Gilstrap (1989) found that mean numbers of O. insidiosus were greatest during anthesis but declined as corn silk became dried and brown during R3. This may have resulted from a lack of fresh corn silk causing unfavorable breeding conditions for the predators during R2 and R3 (Barber 1936). Based on all 3 indices, O. insidiosus exhibited regular or random distributions during R3 of the spring and fall possibly because of low predator density caused by poor breeding conditions or low food supply, but the distributions were otherwise aggregated.

In corn fields of Ohio, Everly (1938) found 6 species of staphylinids including Atheta spp., Barydoma spp., Leptolinus rubripennis Lec., Mycetoporus spp., Philonthus spp., and Coproporus spp., which were collected from corn plants or on nearby ground (Everly 1938). Several staphylinid species are predaceous and possible biocontrol agents of agricultural pests (Cividanes et al. 2009). Larvae of A. insignitus, another predator of ulidiids, were found usually within dry corn silk in corn ears during spring R2 and summer R2 and R3. They were statistically among the most abundant taxa other than ulidiids in the summer R2 and R3 stages, and were numerically highest in the summer R3 (Kalsi et al. 2014). Anotylus insignitus larvae had aggregated distributions in the summer R3, the only season and stage when they were evaluated; hence, they were distributed like the majority of other insect species per corn stage considered.

There were no significant differences in rates of predation by O. insidious among the 3 Euxesta spp. as was also suggested by the similar attack constants and handling times for eggs of the fly species. In the laboratory, O. insidiosus had type III functional responses to eggs of Euxesta spp. A Type III, known as an accelerating functional response, is density dependent and typically exhibited by generalist predators (Murdoch & Oaten 1975). When plotted, a type III functional response has a sigmoid curve and is more characteristic of vertebrates than invertebrate predators or parasitoids (Hassell et al. 1977). However, many invertebrates also have a type III functional response, such as big-eyed bugs feeding on eggs of the cotton bollworm (Shrestha et al. 2004) and o. insidiosus feeding on the mites P. ulmi and T. urticae (Ashley 2003). Possible reasons affecting the occurrence of a type III functional response include selective foraging at high prey density, a predator’s ability to switch prey species, and the effects of learning and experience on prey searching (Holling 1959; Murdoch 1969; Schauber et al. 2004). Orius insidiosus is facultatively phytophagous and can feed on plant material and thus maintain a high population density when prey is scarce (Wiedenmann & O'Neil 1991). Hence, when ulidiid eggs are in clusters that are randomly distributed in tightly packed corn silk, a Type III functional response may also occur because of the availability of plant material and possibly other prey species such as mites. Ashley (2003) reported that hemipteran predators such as O. insidiosus limit their search area by selective foraging, i. e., high frequency turning movements followed by feeding until they reach a threshold time. Also, when prey are scarce, predators may straighten their paths to reach and exploit areas with high prey density. Orius insidiosus may be able to easily feed on eggs in clusters without making many movements, but when the eggs are depleted, it can switch to other prey such as thrips.

The present study found indices of dispersal indicating spatial distributions for ulidiid eggs, larvae, and 2 key predators, O. insidious and to a lesser extent A. insignitus. In addition, handling times, attack constants, and functional responses of O. insidiosus to ulidiid eggs were found. These findings may be compared to those of other predatory arthropod species, such as Z. longipes, and can help in understanding the effectiveness of O. insidiosus and other predators in controlling ulidiid flies. This would permit the further identification and evaluation of potential for predatory arthropods to biologically control all ulidiid life stages.