Food Preference of Pecky Rice Bugs Is Influenced by Experience

This study investigated whether prior experience influences the plant food preference of Trigonotylus caelestialium and Stenotus rubrovittatus which cause pecky rice grain, by using rice plants and 2 poaceous weeds. In a choice experiment between Digitaria ciliaris and the rice plants, both T. caelestialium and S. rubrovittatus adults that were reared on D. ciliaris plants showed significant initial preference for D. ciliaris over rice. In a choice experiment between Echinochloa crus-galli var. aristata and rice plants, T. caelestialium adults reared on E. crus-galli var. aristata strongly preferred E. crus-galli var. aristata over rice throughout the experiment. However, at and after 24 h, T. caelestialium adults reared on rice showed no food preference, although T. caelestialium initially preferred E. crus-galli var. aristata. In contrast, S. rubrovittatus adults reared on rice showed no preference between E. crus-galli var. aristata and rice. However, S. rubrovittatus reared on E. crus-galli var. aristata initially preferred E. crus-galli var. aristata to rice, with this preference waning with time. The same results were obtained for both sexes. Although the effect of experience differed with food source and the species of mirid bug, prior experience initially had a noticeable effect, which disappeared with time (1 d).


Introduction
Phytophagous insects feed on multiple plant species, whereas monophagous insects exhibit high host specificity.
Phytophagous insect preferences are influenced by various factors, especially previous experience [1] [2].The influence of pre-and early-adult experience on host selection has been extensively studied.Although adults are expected to prefer host species that they have been previously exposed to, both positive and negative results have been reported [3] [4].Therefore, the effect of experience requires consideration when investigating host selection by phytophagous insects.
The rice leaf bug, Trigonotylus caelestialium (Kirkaldy), and the sorghum plant bug, Stenotus rubrovittatus (Matsumura) are major rice pests in Japan.Although the nymphs of these bugs grow well on hulled rice, rice plant leaves are not suitable food for their growth.Thus, adults are assumed to cause damage to rice grains when they move to rice plants at the heading stage after growing on other poaceous plants, along with nymphs that emerge from eggs deposited by the adults on rice.Therefore, it is important to identify host plants that are suitable for nymph growth.Field sampling and observations have identified Italian ryegrass, Lolium multiflorum; annual bluegrass, Poa annua; and southern crabgrass, Digitaria ciliaris as important food sources for rice bugs [5]- [7].Laboratory tests have shown that nymphs grow well on P. annua, Alopecurus aequalis, Poasphondylodes, Hordeum murinum, Dactylis glomerata, Agrostis clavata subsp.matsumurae, L. multiflorum [8], Poa pratensis, and Digitaria violascens [9].These plants may facilitate the development of rice bugs.Alternatively, the relative preference of rice bugs between host plants and rice plants may contribute to the occurrence of pecky rice, because rice grain damage is caused by rice bugs that move from poaceous weeds to rice plants.
It has been reported that rice leaf bugs and sorghum plant bugs are attracted to plants at specific developmental stages [10] [11].Furthermore, rice leaf bugs prefer flowering poaceous weeds over rice plants before heading, whereas they show equal preference for flowering poaceous weeds and rice plants at the heading stage [12].Therefore, the increase in bug populations in rice paddies after heading [5] [6] [13] is attributable to an increase in their relative preference for rice plants over other host plants.Differences in the relative preference may affect bug movement between plants.Thus, in addition to the increased effects of poaceous weeds, the relative difference in bug preference between weeds and rice plants must be considered when evaluating how plants affect the occurrence of pecky rice.Moreover, if rice bugs develop a preference for rice plants after experiencing poaceous weeds, these weeds may have a pronounced effect on the incidence of pecky rice.
To study whether bug development on poaceous weeds influences their preference for rice plants, the relative preference between poaceous weeds and rice plants was investigated for T. caelestialium and S. rubrovittatus adults that had had prior experience on different plants.In this study, Digitaria ciliaris and Echinochloa crusgalli var.aristata were used as poaceous weeds because the ears of these plants emerge during the same season as those of rice plants, and because rice bugs grow relatively better on these plants than on other poaceous weeds that are present during the season [8].

Insects
Trigonotylus caelestialium and Stenotus rubrovittatus were collected from fields at the Hokuriku Research Center, National Agricultural Research Center, Niigata Prefecture, Japan.These insects had been reared on young wheat seedlings under laboratory conditions (16:8 L:D photoperiod, 25˚C) for several years as described previously [14]- [16].

Plants
Digitaria ciliaris and Echinochloa crus-galli var.aristata that were growing wild, and rice, Oryza sativa, that was cultivated in the paddy fields, were collected from the fields at the Hokuriku Research Center.The flowering stages of these plants were used for tests.

Experience Treatment
Within 24 h of emergence, 50 -150 T. caelestialium or S. rubrovittatus adults were released in a plastic rearing cage (34 × 25 × 34 cm).Digitaria ciliaris, E. crus-galli var.aristata, and rice with flowering ears were placed in conical flasks (100 ml) filled with water.Each plant was made available to the adults for 4 d.The plants were renewed daily or every other day.This treatment was conducted under laboratory conditions (16:8 L:D photoperiod, 25˚C).

Choice Preference Test
One ear each of D. ciliaris, E. crus-galli var.aristata, and rice was placed in sample tubes filled with water.Two tubes with D. ciliaris and rice or E. crus-galli var.aristata and rice were placed diagonally at the corner of a plastic rearing cage (34 × 25 × 34 cm).Ten adults reared on these plants as described above were released in a plastic cup (top diameter: 8 cm, bottom diameter: 6 cm, height: 4 cm).The plastic cup was placed at the center of the plastic rearing cage and then the cap was removed to allow the adults to disperse.Adults reared on D. ciliaris were released in D. ciliaris vs. rice cages.Those on E. crus-galli var.aristata were released in E. crusgalli var.aristata vs. rice cages.Those on rice were released in D. ciliaris vs. rice cages or E. crus-galli var.aristata vs. rice cages.The number of adults on the plants was counted at 3, 6, 24, 27, and 30 h after starting the test.The test was conducted separately for males and females to prevent interaction between the sexes from influencing food preference.Twelve cages (i.e., 120 individuals) were used for each test, except for T. caelestialium males reared on E. crus-galli var.aristata (7 cages).These tests were conducted under rearing conditions (16:8 L:D photoperiod, 25˚C).

Statistical Analysis
The numbers of adults on rice plants and other poaceous plants were compared at each time point using binomial test.The ratios of the adults present on the plant species on which they had been reared versus those present on the plant species on which they had not been reared were compared between rice and each poaceous species at each time point using Fisher's exact test.For example, at 3 h, the total number of T. caelestialium females reared on D. ciliaris that were present on D. ciliaris and rice was 64 and 25 individuals, respectively (Figure 1(a)).At the same time period, the total number of T. caelestialium females reared on rice that were present on D. ciliaris and rice was 44 and 41 individuals, respectively (Figure 1(b)).These two ratios (64/25 and 44/41) were then compared statistically by Fisher's exact test.If these ratios were significantly different, prior food experience probably influenced bug preference.

Discussion
Stenotus rubrovittatus exhibited different preferences for the two tested poaceous plants (D. ciliaris and E. crus-galli var.aristata) versus rice.In the D. ciliaris-rice combination, after rearing of individuals on D. ciliaris, more S. rubrovittatus adults preferred D. ciliaris to rice.In contrast, after rearing of individuals on rice, no preference between rice and D. ciliaris was initially detected, although more adults frequented D. ciliaris over time (about 1 d).In contrast, for the E. crus-galli var.aristata-rice combination, after rearing of S. rubrovittatus on rice, adults exhibited no preference between E. crus-galli var.aristata and rice at any of the time points (except females at 3 h).However, after rearing of individuals on E. crus-galli var.aristata, more adults initially preferred E. crus-galli var.aristata over rice.This observation contrasted with the results obtained for D. ciliaris, wherein the difference between D. ciliaris and rice disappeared with time.Significant differences in the ratios were initially obtained between plants used for rearing for both the D. ciliaris-rice combination and the E. crusgalli var.aristata-rice combination, with this difference disappearing over time (about 1 d).These results indicate that rearing plant experience influenced food preference initially, but this effect disappeared over time (about 1 d).Variation in the preference of rice bugs for rice and the two weed species (D. ciliaris and E. crusgalli var.aristata) may thus be influenced by innate preferences.
The ratio of the number of S. rubrovittatus adults on D. ciliaris versus rice and E. crus-galli var.aristata versus rice increased with time, possibly because S. rubrovittatus had a stronger preference for D. ciliaris than for E. crus-galli var.aristata.In comparison, the ratio of T. caelestialium adult numbers on D. ciliaris versus rice de-creased with time, whereas it remained consistently higher for E. crus-galli var.aristata versus rice, possibly because T. caelestialium has a stronger preference for E. crus-galli var.aristata than for D. ciliaris.If prior experience affects preference, after rearing of T. caelestialium adults on rice, more individuals would be expected to prefer rice over E. crus-galli var.aristata initially, and then this preference would subsequently decrease.However, the opposite results were obtained.It is possible that the quality of plants in the flasks deteriorated with time.Nevertheless, preference should be influenced by prior experience, because significantly different adult ratios were obtained between the two weeds and rice, depending on the plants used for rearing.Because T. caelestialium had a strong preference for E. crus-galli var.aristata, adults reared on E. crus-galli var.aristata might have retained this preference, even when the test plants deteriorated.In contrast, T. caelestialium adults reared on rice might have selected E. crus-galli var.aristata when the plants were in good condition, switching to rice plants when E. crus-galli var.aristata deteriorated.
Prior experience influenced plant preference behavior of T. caelestialium and S. rubrovittatus.However, the effect of experience disappeared after just 1 d.Experience during specific periods (egg-young larvae) has been shown to influence adult oviposition in a parasitic wasp [17].In contrast, oviposition by Drosophila melanogaster is influenced by odor exposure immediately before testing, with this effect being enhanced by the length of the exposure period [18].Thus, 4 d might have been a too short duration for the rice bugs to have gained sufficient experience in this study.Alternatively, 1 d of experience might be sufficient, and rice bugs might experience and adjust to other plants (i.e., the non-rearing plant) during the tests.Thus, further study is required to understand the relationship between the length of experience and the strength of the effect of experience.
It remains unclear what regulates the relationship between preference and prior experience.Both rice bug species are attracted by odors released from host plants [10]- [12] [19].Attraction behavior is altered by olfactory learning in several phytophagous insects [20]- [24].Therefore, these bugs may be influenced by learning plant odors.However, because the current experiment was conducted in a narrow cage, plant selection by the bugs was probably not caused by their attraction to volatile compounds; rather, they probably remained on the plant following direct contact or sucking (feeding).
Trigonotylus caelestialium and S. rubrovittatus females release sex attractants to attract males [25]- [28].However, it remains unclear how attractants influence the behavior of bugs in narrow cages; thus, the males and females are separated in the selection test to avoid any interaction between sexes.The results showed that both sexes exhibited the same preference trends; thus, differences between the sexes were not considered in this study.
In the current study, adult bugs tended to select plants that they had previously experienced.If the effect of experience in nymphs is sustained to the adult stage, adults may select the same plant species as that on which the nymphs grew.Therefore, if these two rice bug species developed on weeds, they would probably select the same plant species as adults.However, if the weeds deteriorate and disappear, the bugs may transfer and adjust to rice plants in a relatively short period.The results of this study showed that the influence of prior experience declined with time, leading to a situation where the adult ratio on rice plants versus weeds became equal.Thus, these ratios indicate innate preference.Consequently, even though rice bugs grow on weeds, they are expected to adjust to rice plants with time, and spread over rice paddies.
In this study, it was expected that the mirid bugs that developed on poaceous weeds would be inhibited from moving to rice plants because of the effect of food experience, leading to food preference.Although the food preferences of the bugs were influenced by prior experience, this effect was unlikely to be strong.The adult mirid bugs changed their preference for rice plants in a short time period (1 d).In this study, only 2 species of poaceous weeds were used for the tests; however, the two bugs showed different responses to these weeds.Hence, it is important to investigate the effect of other plants.Previous studies have already demonstrated how the oviposition [29] and nymphal growth [8] of these bugs on poaceous weeds may influence the extent to which they may damage rice.However, the factors that influence the relative preference of these bugs for weeds over rice must also be determined, because these bugs develop on poaceous weeds in the vicinity of paddy fields, after which the adults transfer to rice plants.This information would be valuable for understanding how weeds influence rice damage by mirid bugs.

Figure 1 .
Figure 1.Proportion of adult rice bug females on Digitaria ciliaris versus those on rice plants.Choice preference tests between D. ciliaris and rice were conducted with a: Trigonotylus caelestialium reared on D. ciliaris, b: T. caelestialium reared on rice, c: Stenotus rubrovittatus reared on D. ciliaris, and d: S. rubrovittatus reared on rice.The adults were counted at 3, 6, 24, 27, and 30 h.The numbers in parentheses represent the number of adults on plants.† : P values comparing the number of adults on D. ciliaris and that on rice at each time point using binomial test.‡ : P values comparing the proportion of adults present on D. ciliaris and rice (i.e., the plants used for rearing) at the same time points using Fisher's exact test.Asterisk ( * ) represents p < 0.05.

Figure 2 .
Figure 2. Proportion of adult rice bug males on Digitaria ciliaris versus those on rice plants.Choice preference tests between D. ciliaris and rice were conducted with (a): Trigonotylus caelestialium reared on D. ciliaris; (b): T. caelestialium reared on rice; (c): Stenotus rubrovittatus reared on D. ciliaris, and (d): S. rubrovittatus reared on rice.The adults were counted at 3, 6, 24, 27, and 30 h.The numbers in parentheses represent the number of adults on plants.† : P values comparing the number of adults on D. ciliaris and that on rice at each time point using binomial test.‡ : P values comparing the proportion of adults present on D. ciliaris and rice (i.e., the plants used for rearing) at the same time points using Fisher's exact test.Asterisk ( * ) represents p < 0.05.

Figure 4 .
Figure 4. Proportion of adult rice bug males on Echinochloa crus-galli var.aristata versus those on rice plants.Choice preference tests between E. crus-galli var.aristata and rice were conducted with (a): Trigonotylus caelestialium reared on E. crus-galli var.aristata; (b): T. caelestialium reared on rice; (c): Stenotus rubrovittatus reared on E. crus-galli var.aristata, and (d): S. rubrovittatus reared on rice.The adults were counted at 3, 6, 24, 27, and 30 h. † : P values comparing the number of adults on E. crus-galli var.aristata and that on rice at each time point using binomial test.The numbers in parentheses represent the number of adults on plants.‡ : P values comparing the proportion of adults present on E. crus-galli var.aristata and rice (i.e., plants used for rearing) at the same time points using Fisher's exact test.Asterisk ( * ) represents p < 0.05.