Early-Season Host Switching in Adelphocoris spp. (Hemiptera: Miridae) of Differing Host Breadth

The mirid bugs Adelphocoris suturalis (Jakovlev), Adelphocoris lineolatus (Goeze) and Adelphocoris fasciaticollis (Reuter) (Hemiptera: Miridae) are common pests of several agricultural crops. These three species have vastly different geographical distributions, phenologies and abundances, all of which are linked to their reliance on local plants. Previous work has shown notable differences in Adelphocoris spp. host use for overwintering. In this study, we assessed the extent to which each of the Adelphocoris spp. relies on some of its major overwinter hosts for spring development. Over the course of four consecutive years (2009–2012), we conducted population surveys on 77 different plant species from 39 families. During the spring, A. fasciaticollis used the broadest range of hosts, as it was found on 35 plant species, followed by A. suturalis (15 species) and A. lineolatus (7 species). Abundances of the species greatly differed between host plants, with A. fasciaticollis reaching the highest abundance on Chinese date (Ziziphus jujuba Mill.), whereas both A. suturalis and A. lineolatus preferred alfalfa (Medicago sativa L.). The host breadths of the three Adelphocoris spp. differed greatly between subsequent spring and winter seasons. The generalist species exhibited the least host fidelity, with A. suturalis and A. lineolatus using 8 of 22 and 4 of 12 overwinter host species for spring development, respectively. By contrast, the comparative specialist A. fasciaticollis relied on 9 of its 11 overwinter plants as early-season hosts. We highlight important seasonal changes in host breadth and interspecific differences in the extent of host switching behavior between the winter and spring seasons. These findings benefit our understanding of the evolutionary interactions between mirid bugs and their host plants and can be used to guide early-season population management.


Introduction
Host plant use forms the basis of niche breadth and the evolutionary success of herbivores [1]. Depending upon individual host breadth and other ecological particularities, herbivorous insects transfer between host plant species to differing extents to locate suitable host foods for their offspring and themselves [2,3,4,5]. For generalist herbivores, a mixing of diets can produce substantial benefits, and a selective intake of food items or host plant species can redress or prevent nutritional imbalances [6,7]. Subsequently, host switching can be employed as an adaptation to restricted food sources, and eventually result in improved fitness or subsequent population build-up [3,8]. In addition to revealing key aspects of the evolution of plant-animal systems, a knowledge of host breadth and host switching behavior can help to understand the source-sink dynamics of agricultural pests.
In China, Adelphocoris suturalis (Jakovlev), A. lineolatus (Goeze) and A. fasciaticollis (Reuter) are three common pest species on cotton, alfalfa and many other crops [9,10]. Both adults and nymphs feed on the vegetative and reproductive organs of their host plants, causing stunted growth and the abscission or malformation of leaves, flowers and fruits [11]. Over the past 15 years, an increased adoption of transgenic Bt (Bacillus thuringiensis) cotton and the subsequent reduction in insecticide use in this crop have increased Adelphocoris spp. infestation levels [12].
The three Adelphocoris spp. have different geographical distributions, seasonal occurrences and infestation levels. A. suturalis is mainly found in temperate areas, such as the Yangtze River Region and the southern part of the Yellow River Region, whereas A. lineolatus and A. fasciaticollis are usually confined to colder regions, namely certain parts of the Yellow River Region [10,13]. Local agro-landscape composition and the phenology and abundance of suitable host plants are thought to determine Adelphocoris spp. population abundances in each of these regions [9,14]. Each Adelphocoris species has a specific range of overwintering host plants that it uses, largely consistent with each species' distribution and phenology [15].
On these winter hosts, the different Adelphocoris species overwinter as eggs. Some insect species rely on a wide range of plant species for overwintering, whereas others have a far more restricted host range. The eggs of A. suturalis have successfully eclosed from 115 plant species, whereas A. lineolatus and A. fasciaticollis have overwintered on 40 and 35 plant species, respectively [15]. The following spring, the overwintering eggs hatch, and newly emerged nymphs begin feeding on several plant species for one generation; then, the adults subsequently move onto summer host plants. The presence of suitable host plants in or near Adelphocoris spp. overwintering sites is particularly important given the limited dispersal capacities of the newly emerged nymphs [11]. It is unknown to what extent the different Adelphocoris spp. rely on overwinter hosts for spring development and whether early-season host use relates to the dietary breadth of a given species.
In this study, we contrasted the early-season host plant range of the three Adelphocoris species with previously reported winter host use patterns. The results may help explain interspecific differences in the distributions and phenologies of Adelphocoris spp. Additionally, a sound understanding of early-season host switching and population buildup could ultimately help predict Adelphocoris spp. infestation levels in summer crops such as cotton and alfalfa.

Ethics Statement
No specific permits were required for the described field studies.

Field Trials
Field surveys were conducted from mid-April to mid-June of the year 2009-2012 at the natural areas and agricultural fields near the Langfang Experiment Station, Chinese Academy of Agricultural Sciences (CAAS) (116.4 uE, 39.3 uN), in Hebei Province, China. Here, all three Adelphocoris spp. have similarly low population levels [10,15].
Each year, we sampled various plant species (including weeds, fruit trees, economic trees, pastures, and agricultural crops) that are common and widely distributed in the agroecosystems of northern China based on information from local plant guides. A total of 77 plant species from 39 families were sampled, including 53 weeds, 20 trees, 2 pasture crops and 2 agricultural crops. We The sampling protocol was adapted from an existing one [16]. In brief, the Adelphocoris spp. abundance on different plants was assessed using a standard white pan beating method. For herbaceous plants, we examined the entire plant; whereas for tree crops, we sampled the young branches. Sampling was performed every 3-5 days, the plant material was shaken over a 40 cm626 cm611 cm white pan, and the dislodged Adelphocoris individuals (adults and nymphs) were counted [17]. Identification of Adelphocoris species was based on morphological features [18]. Per year, a total of 10-16 sampling events were conducted, with 10-20 random samples taken per plant species and event. For common plant species, a single sample consisted of a total area of 2-20 m 2 , whereas for uncommon species, all of the plants at a given site were sampled. At each event, we determined the exact area covered by each plant species (i.e., sampling area) and recorded the plant growth stage. To correctly identify the associations of a particular Adelphocoris sp. with a given plant species, we only selected uniform patches or carefully chose single stems of a given plant species for sampling. Plant species were identified using regional weed guides [19] or with the assistance of CAAS plant taxonomists. Plant species on which individuals of each Adelphocoris sp. were found were defined as 'host plants' of the respective species [16,20], and those host plants that had a wide distribution and supported high densities of Adelphocoris sp. were regarded as the species' 'dominant hosts' [16].

Statistical Analysis
For each Adelphocoris sp., the average abundance on each plant species was computed on a yearly basis, i.e., by dividing the total number of captured individuals on one plant species by the total area covered by this respective plant throughout the entire sampling period [16]. As the field survey generally started before overwintering eggs had begun to hatch, survey data were not included in the analyses before the appearance of the first individuals of Adelphocoris spp. The abundance of each Adelphocoris sp. was compared between different dominant plant species using a two-way un-replicated ANOVA with a Friedman's test, with years and plant species as fixed factors. A Chi-square test was performed to compare the rate at which overwinter host plants were also used as spring hosts between the three Adelphocoris species. All of the statistical analyses were performed using SAS software [21].

Results
For A. suturalis, 15 species of host plants were found in the spring (Tables 1 and 2), but no significant difference was found for its population abundance on any of the host species (X 2 = 9.  (Tables 1 and 2). On alfalfa, M. sativa, the average abundance of A. lineolatus was 2.7761.21 individuals per m 2 , which was significantly higher than on any other plant (X 2 = 13.16, df = 4, P = 0.0405). The second highest abundance was 0.34 individuals per m 2 on another pasture crop Melilotus suaveolens Ledeb, and those on all of the other host species were less than 0.01. From a total of 12 overwinter host plants, 4 species (incl. the above two pasture crops, and H. scandens, Ziziphus jujuba Mill.) were found to be A. lineolatus's spring host plants (Figure 1b).

Discussion
For Adelphocoris spp., early-season host plants are the key source for future colonization or the exploitation of summer hosts such as cotton. To date, the host plant ranges of various mirid bugs (e.g., Lygus rugulipennis Poppius, Lygus lineolaris (Palisot de Beauvois), Lygus hesperus Knight, Apolygus lucorum (Meyer-Dür)) in the spring have been determined [16,20,22,23,24,25]. This survey determined that there are 15 species of early-season host plants for A. suturalis, 7 species for A. lineolatus, and 35 species for A. fasciaticollis in northern China. Several early-season host plants had been previously identified for these Adelphocoris spp. in China [9,14,26,27,28]. However, because these studies were conducted at different locations with differing species compositions and differing abundances of Adelphocoris spp. and plants, their results cannot be used to explore between-species differences in distribution and seasonal occurrence. Our present study effectively complements previous work because all three Adelphocoris spp. coexist at similar population levels at the study site [15].
In 2008, literature reviews and exploratory host range trials indicated that there was a total of 116, 125 and 30 host plant species for A. suturalis, A. lineolatus and A. fasciaticollis, respectively [11]. Novel work brought the respective host plant range of A. suturalis, A. lineolatus and A. fasciaticollis to 270, 245 and 127 species, maintaining the previous interspecific differences in host breadth (Lu YH, unpublished data; Table 3). Because of the limited abundance/cover at sampling sites, certain plant species were only sampled in 1-2 m 2 in this study. Although limited sampling might lead to underestimates of the host range of a given Adelphocoris spp., plant species with low abundance/cover in natural and agricultural habitats will only play a minor role in the population dynamics of the different mirid bugs. Hence, the updated results presented here provide a comprehensive set of information on year-round host plant range for future research on the interactions Sampling area refers to the combined area covered by each sampled plant species in the respective year. The information of overwinter host ranges of Adelphocoris spp. is cited from [15]. The signs ''+'' and ''2'' indicate that the associated plant species is a host plant or non-host plant, respectively. A blank space indicates that this species was not surveyed. Plant species highlighted in bold were included in the host plant surveys during the winter [15] and the spring (present study). doi:10.1371/journal.pone.0059000.t001 Sampling area refers to the combined area covered by each sampled plant species in the respective year. The information of overwinter host ranges of Adelphocoris spp. is cited from [15]. The signs ''+'' and ''2'' indicate that the associated plant species is a host plant or non-host plant, respectively. A blank space indicates that this plant was not surveyed. Plant species highlighted in bold were included in the host plant surveys during the winter [15] and the spring (present study). doi:10.1371/journal.pone.0059000.t002 between Adelphocoris spp. and its host plants and the regional management of these polyphagous pests. Stark differences were found between the host breadth of the three Adelphocoris spp. during the winter and spring season. For A. suturalis, a limited set of host plants was found during the spring compared to their overall host range and overwinter host range of 270 and 115 species, respectively (Table 3). In the Yangtze River region, where A. suturalis is dominant, several important host plants such as horse bean (Vicia faba L.), carrot (Daucus carota L.), garland chrysanthemum (Chrysanthemum coronarium L.), celery (Apium graveolens L.), alfalfa and hairy vetch (Vicia villosa Roth), are cultivated to a large extent [9,27]. The fact that the above host plants are grown to a lesser extent in northern China may partly explain the relatively low population levels of this pest locally.
For A. lineolatus, alfalfa was the principal early-season host plant and is also an important overwinter host for this species [15]. Large areas of alfalfa, cultivated as a pasture crop, could explain the relatively high population levels of A. lineolatus in one of China's key cattle growing areas (i.e., Cangzhou, Hebei Province) [11,29]. Indeed, A. lineolatus adults greatly prefer alfalfa to other host plants [29], but periodic rotation of alfalfa fields can cause adults to disperse to cotton, sunflower and other crops. As new alfalfa fields are established, A. lineolatus adults gradually migrate back to the alfalfa fields [29]. These phenomena indicate that alfalfa is the most important host plant for A. lineolatus, which greatly affects its distribution and phenology.
For A. fasciaticollis, the early-season host range was similarly as broad as the overwinter host range, with 35 plant species reported as overwinter hosts [15]. Chinese date was the most important overwinter and early-season host plant. It was previously thought that trees were significant hosts for A. fasciaticollis, but because no individuals were found on other fruit trees, such as P. persica and Malus domestica Borkh., the A. fasciaticollis life cycle may be mainly restricted to Chinese date [11,26]. Chinese date is primarily grown in northern China [30], which could explain why A. fasciaticollis is mainly confined to this part of the country [10,18].
During the host-plant selection process of phytophagous insects, the successful colonization of suitable host plants is pivotal for their individual survival and population build-up. For specialist insect species, it may be more difficult and dangerous to change food plants and seek a new host than for generalists [31]. Hence, in general, the degree of host fidelity of comparative specialists tends to be higher than for generalists [32]. In our study, the different Adelphocoris species exhibited varying levels of fidelity to their overwinter host plants, with the (comparative) specialist A. fasciaticollis exhibiting the greatest extent of host fidelity. This finding supports the above general viewpoint on host fidelity of phytophagous insects.
Host fidelity does not necessarily imply increased survival because host switching can cause additional mortality. Even for species that use overwinter host plants for spring development, survival rates can be as low as 30% [33]. For species such as A. suturalis and A. lineolatus, that use an entirely new set of plants for early-season development, host switching could constitute an additional mortality factor [34]. Consequently, it is expected that host switching leads to a fitness increase that effectively compensates for this additional mortality. In addition to host plant ranges, the fitness of Adelphocoris spp. on different hosts can thus help explain between-population differences in many life-history traits.
Our work shows large seasonal variability in host usage patterns. For A. suturalis and A. lineolatus, a relatively small set of host plants was recorded during the spring compared to their overall host range, which comprises 270 and 245 species, respectively (Table 3). A. fasciaticollis adopted a fairly similar host range in spring and winter seasons but exhibited the broadest host range in the spring season, being a comparative specialist. Seasonal differences in host usage likely relate to the nutritional profile of a given plant species for (spring) nymphal development versus physical attributes that provide shelter for winter eggs. Nevertheless, the large differences in host plant ranges of both A. suturalis and A. lineolatus between subsequent seasons must be further analyzed. More precisely, the relationship between (autumn) adult oviposition preference and offspring performance merits further study [35,36]. As both of the populations appear to experience a 'bottleneck' in the spring, important opportunities for population management could be identified [37].
Because Adelphocoris spp. complete their first generation on earlyseason host plants, these plants act as important sources for subsequent infestation of cotton and other summer agricultural crops. Hence, strategic management of early-season host plants could lead to important reductions of those summer populations. For instance, in the United State, broadleaf weeds are the main early-season host plants of the tarnished plant bug L. lineolaris before its movement into cotton fields [38]. Systematic removal of stands of broadleaf weeds near cotton plantings effectively reduced subsequent L. lineolaris numbers in cotton fields [39,40]. Our work provides the basis for similar tactics for the suppression of earlyseason populations of Adelphocoris spp. in cotton agroecosystems in China.