Introduction

Bagrada hilaris Burmeister (Hemiptera: Pentatomidae), commonly referred to as the Bagrada bug or Painted bug, is an invasive stink bug species that has been infesting Brassicaceous crops in the desert southwest of the USA for the past few years (http://cisr.ucr.edu/bagrada_bug.html). Populations of B. hilaris have expanded their range into the coastal cole crop production regions of California and central Arizona, and more recently have been reported in Nevada, New Mexico, and Utah (Bundy et al. 2012). It is considered a serious economic pest of a variety of Brassicaceous vegetable crops grown during the fall and winter months in the desert agricultural valleys of Arizona and southern California (Palumbo and Natwick 2010). Seedling stages of Brassicaceous crops appear to be highly susceptible to direct feeding damage on cotyledons, newly emerged leaves, and apical meristems. Excessive feeding damage to apical meristems may result in destruction of the terminal growing points, leading to either adventitious bud break (e.g., cabbage plants with multiple, and unmarketable heads) or plants with no marketable heads being formed (e.g., broccoli with no crowns) (Palumbo and Natwick 2010). In the three years since the initial outbreaks in 2009 of B. hilaris on desert vegetable crops, it has become evident that the invasive stinkbug has become an established pest of commercial Brassicaceous crops. Recent surveys of broccoli growers from Yuma, AZ, and the Imperial Valley, CA estimated that about 90 % of the broccoli acreage planted since 2010 have been infested with B. hilaris at some point in the growing season, which on average has resulted in considerable stand losses and plant injury, often exceeding 10 % yield losses. (https://extension.arizona.edu/sites/extension.arizona.edu/files/resourcefile/resource/marcop/052913%20Bagrada%20Bug%20Survey_2013_Report.pdf). Furthermore, the potential economic impact of B. hilaris on the western vegetable industry could be significant considering that the production of broccoli, cauliflower, cabbage, and other Brassicaceous crops in Arizona and California was collectively valued at over $1 billion in 2011 (CDFA 2012; USDA-NASS 2011, 2012).

Limited information is available on the impact of feeding by B. hilaris on Brassicaceous vegetables. Recently, field and laboratory observations demonstrated the damage potential of B. hilaris infesting broccoli in Arizona (Palumbo and Natwick 2010; Huang et al. 2013). Plant responses of broccoli to herbivory have been quantified for a related stink bug species, the harlequin bug, Murgantia histrionica Hahn (Ludwig and Kok 2001), but scientific information is lacking on the direct effects of feeding damage by B. hilaris on Brassica host plants and associated growth responses. In other studies, reductions in plant height, dry root, and leaf weights on seedling corn were reported by the feeding of three pentatomids, the brown stink bug, Euschistus servus Say, one-spotted stink bug, E. variolarus Palisot de Beauvois, and the green stink bug, Chinavia hilaris Say (Townsend and Sedlacek 1986). Negative plant growth responses as influenced by feeding of other hemipterans has been studied in the grass bug, Irbisia pacifica Uhler (Hansen and Nowak 1988), silverleaf whitefly, Bemisia argentifolii Bellows & Perring (Palumbo et al. 2000), a phloem-sap feeding aphid, Uroleucon caligatum Richard, and a xylem-sap feeding meadow froghopper, Philaenus spumarius (L.) (Meyer and Whitlow 1992).

Feeding and oviposition by two phytophagous pentatomids, the southern green stink bug, Nezara viridula (L.), and M. histrionica were shown to negatively affect photosynthesis of French bean, Phaseolus vulgaris, and savoy cabbage, Brassica oleracea (Velikova et al. 2010). Losses of chlorophyll content in leaves in response to feeding by other hemipterans have been reported on the sweetpotato whitefly, Bemisia tabaci Gennadius (Buntin et al. 1993; Palumbo et al. 1996), azalea lace bug, Stephanitis pyrioides Scott (Buntin et al. 1996), Russian wheat aphid, Diuraphis noxia Mordvilko (Ni et al. 2002; Heng-Moss et al. 2003), black pecan aphid, Melanocallis caryaefoliae Davis (Cottrell et al. 2009), and pea aphid, Acyrthosiphon pisum Harris (Goławska et al. 2010). Similar reductions in leaf area and relative chlorophyll content resulting from B. hilaris feeding have not been previously documented.

A clear understanding of plant health and growth responses associated with the feeding injury on host crops caused by B. hilaris will be important for the development of pest management approaches for this new pest. This is particularly important considering that recent field observations showed that B. hilaris adults are often attracted in large numbers to seedling stages of broccoli during crop establishment (Huang et al. 2013). The main objectives of this study were to quantify feeding damage of B. hilaris on cotyledon, 2-leaf, and 4-leaf stages of six commercial Brassicas and to evaluate host plant responses to B. hilaris feeding intensity by measuring leaf growth, relative chlorophyll content, and dry weight in a series of laboratory trials.

Materials and methods

Insects and plants

Commercial Brassicaceous vegetable crops including broccoli (B. oleracea var. italica Plenck ‘Emerald Crown,’ Sakata Seed, Morgan Hills, CA), green cabbage (B. oleracea var. capitata L. ‘Gazelle F1,’ Bejo Seeds, Oceano, CA), red cabbage (B. oleracea var. capitata L. ‘Ruby Perfection Hybrid,’ American Takii, Inc., Salinas, CA), cauliflower (B. oleracea var botrytis L. ‘Ponderet RZ Hybrid,’ Rijk Zwaan USA Inc, Salinas, CA), kale (B. oleracea var. acephala DC.’Winterbor Hybrid,’ Terriorial Seeds, Cottage Grove, OR), and radish (Raphanus sativus var. sativus L. ‘Rover Hybrid’) were selected as test crops based on previous choice experiments that showed these host plants to be preferentially attractive to B. hilaris (Huang et al. 2014). All plants were direct-seeded into 5 × 5 cm2 pots for germination with potting soil (Miracle-Gro®, Marysville, OH) and irrigated daily in the greenhouse. Once plants reached the 2-leaf stage, they were fertilized weekly with an all-purpose water-soluble fertilizer (20-20-20) (Miracle-Gro®). No fertilizer was applied to plants at the cotyledon stage. Plants of uniform size at the cotyledon stage (6–8 day-old), 2-leaf stage (20–22 day-old), and 4-leaf stage (~35 day-old) were selected for the no-choice experiment.

All adult insects used for the experiment were obtained from a B. hilaris colony maintained at the Yuma Agricultural Center, Yuma, AZ, where all life stages were supplemented with organic broccoli heads (Earthbound Farm, Salinas, CA) and sweet alyssum plants. Single adult B. hilaris (females), one mating pair (sexually mature), and two mating pairs were chosen and starved for 24 h before being caged on plants used in the studies.

No-choice tests

A 4 × 3 factorial design experiment was conducted in an upright growth chamber (model I-36 LLVL, Percival Scientific, Perry, IA) by caging B. hilaris adults (zero, a single female, one mating pair, and two mating pairs) on individual host plants (cotyledon, 2-true leaf, and 4-true leaf stages). Environmental condition was maintained a photoperiod of 12:12 (L:D) h and temperature at 30 °C during the day and 20 °C at night; these conditions were confirmed and monitored using a data logger (HOBO Pendant, Onset Computer Corp., Cape Cod, MA). Individual potted host plants were transferred into small cages constructed from an 8.5-cm-diameter plastic cup and placed into the growth chamber 24 h before the experiment. The top of the plastic cup was covered by a 9-cm-diameter Petri-dish cap drilled with 15–20, 0.2-cm-diameter round holes for ventilation. B. hilaris adults (a single female, one mating pair, and two mating pairs) were caged onto separate plants, and no adults were caged on the control group plants. Each combination (e.g., two mating pairs on 2-true leaf-stage plant) was replicated six times. B. hilaris were allowed to feed on the plant for 48 h, after which time all adults were removed from the plants. Leaf area measurements for each plant were recorded immediately before the experiment (24 h in the chamber) and again following the 48-h infestation period. Nondestructive estimates of total leaf area were conducted using a USB digital microscope (Dino-Lite®) and measured using Dino Capture 2.0 graphic software. Leaf growth/loss was determined based on positive or negative changes in leaf area after the 48-h feeding period. Feeding damage on leaf tissue, as described by Palumbo and Natwick (2010), was measured after the 48-h period similar to methods used in a previous study (Huang et al. 2013). Relative chlorophyll content in leaf tissue was determined immediately before and after the 48-h infestation period using a nondestructive chlorophyll meter (SPAD-502 Plus, Minolta Corp., Spectrum Technologies, Inc.). Two measurement points were randomly selected on both sides of each cotyledon, four points were selected on both sides of each leaf for the 2-leaf groups, and six sample points were measured on both sides of all leaves for the 4-leaf-stage plants. SPAD (Soil and Plant Analyzer Development) values obtained from each plant were averaged for further analyses. Chlorophyll change was based on positive or negative changes in SPAD values after the 48-h period of B. hilaris infestation.

All 4-leaf-stage plants were transferred into the greenhouse following the 48-h infestation period and allowed to continue to grow for 21 days to measure residual impacts of feeding on plant growth. After the 21-day period, plants were abscised 1 cm above soil surface and total leaf area was calculated by measuring the leaf area of all leaves on each plant with a LI-3100 leaf-area meter (LI-COR, Inc., Lincoln, NE). Dry weights for each plant were measured by drying all leaves, petioles, and stems in a forced air oven held at 150 °C for 24 h. Dried bulb weight (underground portion) was also recorded in radish plants using the same method as dry plant weight.

Statistical analyses

This study was designed to measure the effects of plant stage and numbers of B. hilaris per plant for each Brassica host rather than comparisons among hosts. Thus, data obtained from each Brassicas host were analyzed separately using a two-way ANOVA (PROC GLIMMIX; SAS Institute 2009). Plant stage and number of B. hilaris infested were modeled as fixed effects, and replicates were modeled as a random effect. The response variables, such as percentage of feeding damage (damaged area/total leaf area), were subjected to arcsine square root transformation before analysis (Zar 1999). Actual untransformed data are presented in the tables and figures. Treatment means were separated using the LSMEANS test (P < 0.05) and the slice option (SAS Institute 2009) if there was a significant interaction. A t test (PROC TTEST, SAS Institute 2009) was used to determine whether there were differences in feeding damages between upper two leaves and lower two leaves of 4-leaf-stage plants. Analyses of dry weight and total leaf area (4-leaf stage only) were performed using a one-way ANOVA (PROC GLIMMIX; SAS Institute 2009) to compare the difference between number of B. hilaris infested. Relationship between number of B. hilaris infested and response variables was quantified using regression analysis (PROC REG) and Spearman’s rank correlation coefficient. A paired t test was conducted to compare the changes (before and after B. hilaris infestation) in chlorophyll content in each host.

Results

Feeding damage

When examining the percentage of feeding damage caused by B. hilaris, there was a significant interaction between plant stage and number of B. hilaris infested in each Brassicaceous host: broccoli (two-way ANOVA, F 9,75 = 17.27, P < 0.0001), green cabbage (two-way ANOVA, F 9,75 = 12.53, P < 0.0001), red cabbage (two-way ANOVA, F 9,75 = 5.94, P < 0.0001), cauliflower (two-way ANOVA, F 9,75 = 27.26, P < 0.0001), kale (two-way ANOVA, F 9,75 = 13.27, P < 0.0001), and radish (two-way ANOVA, F 9,75 = 15.20, P < 0.0001). Further analyses of the interaction, sliced by plant stage and by number of B. hilaris infested, are shown in Table 1. When caged with one female, one mating pair, or two mating pairs of B. hilaris, the percentage of feeding damage varied significantly among different growth stages in all hosts (Table 1). In addition, the percentage of feeding damage varied significantly among different numbers of B. hilaris infested in cotyledon, 2-leaf stage, and 4-leaf (upper two leaves) stages of all hosts (Table 1). In general, the percentage of feeding damage increased as the numbers of B. hilaris increased; this was true in most of the cotyledon and 2-leaf-stage plants from all hosts (Fig. 1). Percentage feeding damage was generally lower in the 4-leaf-stage plants compared to the cotyledon and 2-leaf-stage plant stages, but in most cases, damage was significantly greater on infested plants compared to the non-infested control (Fig. 1). However, regardless of number of B. hilaris infested, estimates of feeding damage on the upper (younger) two leaves were significantly greater than those measured on the lower (older) two leaves of the 4-leaf-stage plants in broccoli (t test, t = 6.28, P < 0.0001), green cabbage (t test, t = 5.67, P < 0.0001), red cabbage (t test, t = 6.31, P < 0.0001), cauliflower (t test, t = 7.84, P < 0.0001), kale (t test, t = 6.01, P < 0.0001), and radish (t test, t = 6.37, P < 0.0001). Regression analyses showed a positive relationship between feeding damage and number of B. hilaris infested in cotyledon, 2-leaf, and 4-leaf (upper two leaves) stages, but not in the 4-leaf (lower two leaves) stage (Table 2).

Table 1 Analysis of the interaction between plant growth stage and number of B. hilaris infested in percentage of feeding damage using a two-way ANOVA with slice option for each host plant
Full size table
Fig. 1
figure 1

Mean ± SEM percentage leaf damage of six Brassica crops exposed to various numbers of B. hilaris adults for 48 h on seedling plants at a cotyledon stage, b 2-leaf stage, c 4-leaf stage, younger two leaves, and d 4-leaf stage, older two leaves. Leaf damage was based on the direct measurement of the chlorotic leaf tissue induced by feeding of B. hilaris. Percentage of leaf damage was based on the chlorotic leaf area/total leaf area. Means followed by the same letter are not significantly different (P < 0.05; LSMEANS test)

Full size image
Table 2 Relationship between feeding damage and number of B. hilaris infested at different growth stages of Brassica plants
Full size table

Leaf area growth/loss

When examining the growth/reduction in total leaf area as a response to B. hilaris feeding, there was a significant interaction between plant stage and number of B. hilaris infested in broccoli (two-way ANOVA, F 9,75 = 21.53, P < 0.0001), green cabbage (two-way ANOVA, F 9,75 = 7.45, P < 0.0001), red cabbage (two-way ANOVA, F 9,75 = 5.76, P < 0.0001), cauliflower (two-way ANOVA, F 9,75 = 4.62, P < 0.0001), kale (two-way ANOVA, F 9,75 = 2.13, P = 0.0370), and radish (two-way ANOVA, F 9,75 = 8.63, P < 0.0001). Further analyses of these interactions, sliced by plant stage and by number of B. hilaris infested, are shown in Table 3. When caged with zero adult, one female, one mating pair, or two mating pairs of B. hilaris, changes in total leaf area varied significantly among the different growth stages in all hosts (Table 3). In addition, leaf growth/loss varied significantly among different numbers of B. hilaris infested, except for cotyledon stage of green cabbage, red cabbage, and 4-leaf stage (lower two leaves) of red cabbage, cauliflower, and kale (Table 3). In the control groups, leaf area increased in all stages within a 48-h period. In general, the leaf growth decreased as the numbers of B. hilaris increased, especially in cotyledon and 2-leaf stages (Fig. 2). Regression analyses showed a negative relationship between leaf growth and number of B. hilaris infested; this was especially obvious in cotyledon and 2-leaf-stage plants from all hosts (Fig. 2; Table 4).

Table 3 Analysis of the interaction between plant growth stage and number of B. hilaris infested in leaf growth/loss using a two-way ANOVA with slice option for each host plant
Full size table
Fig. 2
figure 2

Mean ± SEM change in leaf growth (cm2) of six Brassica crops exposed to various numbers of B. hilaris adults for 48 h on seedling plants at a cotyledon stage, b 2-leaf stage, c 4-leaf stage, younger two leaves, and d 4-leaf stage, older two leaves. Leaf growth was based on positive or negative changes in leaf area after a 48 h of B. hilaris infestation. Means followed by the same letter are not significantly different (P < 0.05; LSMEANS test)

Full size image
Table 4 Relationship between leaf growth and number of B. hilaris infested at different growth stages of Brassica plants
Full size table

Change in relative chlorophyll content

When examining the relative chlorophyll content in plant leaves in response to B. hilaris feeding, there was a significant interaction between plant stage and number of B. hilaris infested in broccoli (two-way ANOVA, F 9,75 = 2.72, P = 0.0085), green cabbage, (two-way ANOVA, F 9,75 = 3.33, P = 0.0018), red cabbage (two-way ANOVA, F 9,75 = 2.49, P = 0.0152), cauliflower (two-way ANOVA, F 9,75 = 6.02, P < 0.0001), kale (two-way ANOVA, F 9,75 = 3.12, P = 0.0031), and radish (two-way ANOVA, F 9,75 = 4.48, P = 0.0001). Further analyses of the interactions, sliced by plant stage and by number of B. hilaris adults infested, are shown in Table 5. When caged with one female, one mating pair, or two mating pairs of B. hilaris, the relative chlorophyll change varied significantly among different growth stages in all hosts, but not in the controls (Table 5). The relative chlorophyll content varied significantly among different numbers of B. hilaris infested in cotyledon, 2-leaf, and 4-leaf (upper two leaves) plants, but not in the 4-leaf (lower two leaves) plants (Table 5). Regression analyses showed a negative relationship between relative chlorophyll change and number of B. hilaris infested (Table 6). In general, the SPAD value decreased with the increase in B. hilaris infested in cotyledon, 2-leaf, and 4-leaf (upper two leaves) from all hosts (Fig. 3). However, the SPAD values measured at 4-leaf (lower two leaves) was not influenced by the increase in B. hilaris infested. There was a negative correlation between relative chlorophyll change and amount of feeding damage in all hosts (Spearman’s coefficient: broccoli = − 0.6937; green cabbage = −0.7298; red cabbage = −0.7043; cauliflower = −0.7671; kale = −0.7561; and radish = −0.7895).

Table 5 Analysis of the interaction between plant growth stage and number of B. hilaris infested in chlorophyll content (SPAD value) using a two-way ANOVA with slice option for each host plant
Full size table
Table 6 Relationship between change in relative chlorophyll content and number of B. hilaris infested at different growth stages of each Brassica host plant
Full size table
Fig. 3
figure 3

Mean ± SEM change in leaf chlorophyll content (SPAD values) of six Brassica crops exposed to various numbers of B. hilaris adults for 48 h on seedling plants at a cotyledon stage, b 2-leaf stage, c 4-leaf stage, younger two leaves, and d 4-leaf stage, older two leaves. Chlorophyll change was based on positive or negative changes in SPAD value after a 48-h period of B. hilaris infestation. Means followed by the same letter are not significantly different (P < 0.05; LSMEANS test)

Full size image

Dry weight and total leaf area in 4-true leaf-stage plants

There were significant differences in both total leaf area and dry plant weight between the different numbers of B. hilaris infested in broccoli, green cabbage, red cabbage, cauliflower, and radish, but not in kale (Table 7). Dry plant weight and total leaf area were greatest in the control groups, and lowest in plants infested with the groups of 2-mating pairs of B. hilaris. In radish, the dry bulb weight varied significantly among the different B. hilaris infestation levels (one-way ANOVA, F 3,20 = 6.98, P = 0.0021), where control groups had the greatest bulb weight and groups of 2-mating pairs had the lowest bulb weight. Spearman’s correlation coefficient analyses showed a negative correlation between leaf area/dry plant weight and number of B. hilaris infested. These variables decreased as the number of B. hilaris increased, with kale as the only exception (Table 7).

Table 7 Correlation between number of B. hilaris infested and mean ± SEM total leaf area and dry weight from 4-leaf-stage plants 21 days following the infestation period for each host plant
Full size table

Discussion

Our results clearly demonstrate that feeding by adult B. hilaris can cause negative impacts on the leaf growth, chlorophyll content, and dry weights of several important Brassicaceous host plants. Feeding damage measured on individual host plants varied significantly, depending on adult density and stage of plant growth. Feeding was particularly destructive to the younger developmental plant stages such as cotyledon and 2-true leaf-stage plants. In general, 4-true leaf-stage plants were less vulnerable to B. hilaris feeding in all the host plants evaluated. Differences in feeding damage measured between the upper (younger) two leaves and lower (older) two leaves of the 4-leaf-stage plants further showed that B. hilaris fed preferentially on younger growing tissue. Furthermore, feeding damage on the upper two leaves (younger) caused significant residual growth effects on the 4-leaf-stage plants. Although the study was not designed to compare differences among the six Brassicaceous host plants, it was quite evident that older kale plants at the 4-true leaf stage were less susceptible to B. hilaris feeding compared to other 4-leaf host plants based on the minimal differences in total leaf area and dry weight 21 days after the 48 h of B. hilaris infestation period.

In addition to plant stage, negative impacts on leaf area reduction and relative chlorophyll loss were significantly related to the number of B. hilaris infested. Relative chlorophyll loss was also related to the amount of feeding damage caused by B. hilaris. These data are consistent with anecdotal observations and empirical studies (J. C. Palumbo, unpublished) that have shown that heavy infestations of B. hilaris occurring in seedling broccoli and cauliflower fields can cause significant plant injury and yield reductions. Individual females were used for the 1-adult group rather than males in our study because previous research revealed that female B. hilaris fed for significantly longer durations and caused almost fivefold more feeding damage than males (Huang et al. 2013). Another related species, M. histrionica, was reported to cause plant mortality in broccoli plants more quickly as insect density increased (Ludwig and Kok 2001).

Brown et al. (2003) showed that glucosinolates, the secondary plant-defensive compound in the family of Brassicaceae, were detected at higher concentration in younger leaves than older leaves. Preferential feeding of B. hilaris on younger leaf tissue in 4-leaf-stage plants suggests that putatively higher concentrations of glucosinolates caused no significant reduction in B. hilaris feeding. Gols and Harvey (2009) reported that insect herbivore specialists on Brassicaceous plant species have evolved adaptations to excrete or detoxify glucosinolates. The flea beetle, Phyllotreta cruciferae Goeze, did not discriminate between cotyledons having sinigrin [major glucosinolate of Brassica juncea (L.) Czern.] and glucobrassicin (major glucosinolate of B. napus L.), and both glucosinolate types, and concentrations had little effect protecting seedlings of these two Brassica from P. cruciferae feeding damage (Bodnaryk and Palaniswamy 1990). These relationships need to be explored further for B. hilaris and susceptible Brassicaceous hosts.

The impact of insect feeding often varies by species. For instance, insects such as aphids can feed without affecting plant growth or fitness, while feeding by other insects such as spittlebugs and beetles can reduce vegetative growth and seed production (Meyer and Whitlow 1992; Meyer 2000). In this study, we observed that leaf growth was negatively affected by the feeding of B. hilaris. Significant reductions in total leaf area at the cotyledon and 2-leaf stages for all tested host crops indicated that early developmental plant stages are very sensitive to B. hilaris feeding. Other hemipterans have been shown to reduce leaf area following feeding bouts. The grass bug, I. pacifica, was reported to cause a similar reduction in both green and total leaf area on Great Basin wildrye and intermediate wheatgrass (Hansen and Nowak 1988). Feeding by Bemisia whiteflies was demonstrated to cause negative impacts on alfalfa growth, forage yield, and quality (Palumbo et al. 2000). Rice bug, Leptocorisa oratorius (F.), was reported to reduce grain quality, seed germination rates, and yield loss in the Philippines (Jahn et al. 2004).

The chlorophyll meter used in this study provided a nondestructive, quick, and reliable method for measuring chlorophyll changes associated with insect feeding damage. In our study, chlorophyll content in host plant leaf tissue was negatively affected even when plants were only exposed to a single female. As anticipated, no significant decrease in chlorophyll content was detected in control groups because of the lack of feeding damage. Relative SPAD values did, however, fluctuate among host plants as would be expected. Although different parameters were measured, Velikova et al. (2010) reported a significant reduction in chlorophyll fluorescence parameters on herbaceous plants resulting from the feeding and oviposition of two pentatomids, M. histrionica and N. viridula. They confirmed that the permanent impairment of photosynthetic phytochemistry was restricted to the damaged areas on the leaf (Velikova et al. 2010).

Losses of chlorophyll content due to insect herbivory have been quantified in other hemipterans using similar chlorophyll analyses. Feeding by B. tabaci was reported to reduce relative leaf chlorophyll levels (SPAD values) in lettuce leaves (Palumbo et al. 1996). Feeding by A. pisum caused significant loss of chlorophyll a and b in the Fabaceae plants (Goławska et al. 2010). Infestation of D. noxia was reported to be associated with the loss of chlorophyll content in cereal crops using a chlorophyll meter (Heng-Moss et al. 2003). Measurements with a Spectronic 401(Milton Roy, Rochester, NY) showed that feeding injury by S. pyrioides reduced chlorophyll content and adversely affected net leaf photosynthesis and transpiration in azalea (Buntin et al. 1996). B. tabaci was reported to reduce leaf photosynthesis in tomato leaves by reducing content and photosynthetic capacity of chlorophyll (Buntin et al. 1993).

Residual dry weights and total leaf areas for the host plants 21 days following the infestation period were not available for the cotyledon and 2-leaf stages in our experiment. Because each leaf was detached from the plant at the end of the 48-h period for measurements of leaf area and feeding damage under the digital microscope, these plants could not be used. However, 4-leaf plants were large enough that a nondestructive method was used to measure leaf area and damage without removing the leaves from the plant (Huang et al. 2013). Although yields of host plants were not examined in this study, increases in total leaf area by greater than tenfold following the 21-day growth period in the greenhouse suggested that 4-leaf-stage plants may be able to compensate for short bouts of feeding by B. hilaris. However, the significant reductions in dry plant weight (above ground) and dry bulb weight (underground) in the damaged radish plants indicated that feeding of B. hilaris can likely reduce yields and market value of radish crops under field conditions. Reductions in yield and quality by feeding of pentatomids were reported in cotton (Barbour et al. 1990; Greene et al. 1999) and vegetables (Ludwig and Kok 2001; Kuhar et al. 2012). Townsend and Sedlacek (1986) reported that feeding by C. hilaris and Euschistus species produced significant detrimental effects to seedling and early development stages of corn, including reductions in plant height, dry root, and leaf weights.

Our data provide essential information on the impact of B. hilaris feeding on plant growth and chlorophyll for several commercial Brassicaceous vegetables. Losses in total leaf area, chlorophyll content, and dry weight in response to short bouts of B. hilaris feeding suggest a feeding-induced stress response in early developmental stages of Brassicaceous species. Because of the economic significance and large number of acres of Brassicaceous crops grown in the desert southwest, control of this pest during crop establishment is critical. To date, insecticides have been the primary method for B. hilaris management. Action thresholds have not yet been established for B. hilaris on these crops largely due to the lack of data describing density/damage relationships between this pest and commercial cole crops. Data generated in this study have provided a better understanding of the impacts of B. hilaris feeding damage on seedling Brassicaceous plants and should be useful in developing pest management strategies during crop establishment. Field studies are presently underway to examine which phenological growth stages of seedling broccoli and cauliflower plants (i.e., cotyledon through 6-leaf-stage plants) can tolerate or compensate B. hilaris feeding damage without suffering economic reductions in yield, quality, and harvest maturity.