Volatiles Induced from Hypolepis punctata (Dennstaedtiaceae) by Herbivores Attract Sclomina erinacea (Hemiptera: Reduviidae): Clear Evidence of Indirect Defense in Fern

Simple Summary Plants have developed many kinds of indirect defense mechanisms against herbivorous insects, through which the natural enemies are attracted by herbivore-induced plant volatiles to prey on these phytophagous species. There have been many reports regarding this mechanism in spermatophytes, but little is known in ferns. In this study, the relationship between the fern species Hypolepis punctata and the predatory insect Sclomina erinacea was studied. Employing field observations, plant volatile analysis, electrophysiological experiments, and behavioral experiments, we found that not only the plant can attract the assassin bug in the field, but compared with the healthy individuals, the fern being damaged by herbivorous insects also emitted several novel volatile organic compounds, which were strongly attractive to the assassin bug. The results indicate that ferns may also have indirect defense mechanisms using volatile organic compounds, and it is the first report of an indirect defense mechanism in fern. Abstract Plants have evolved various self-defense mechanisms against insect feeding. There are many reports regarding both direct and indirect defense mechanisms in seed-plant. However, only direct defenses on ferns were considered and the indirect defense mechanism has never been reported. In this study, it was observed that the fern Hypolepis punctata can attract the assassin bug Sclomina erinacea in the field. We collected and analyzed volatiles from H. punctata healthy individuals and the ones wounded by Bertula hadenalis, using dynamic headspace and GC-MS. We recorded the electroantennogram responses of antennae of S. erinacea to different standards of volatile compounds identified from the GC-MS analysis. We also analyzed the behavior of male and female S. erinacea adults in response to volatiles collected from H. punctata using a Y-tube olfactometer. The results showed that a number of volatile compounds were produced when the fern was damaged by B. hadenalis. Electroantennography and Y-tube olfactometer results showed that some herbivore-induced volatiles and volatiles from undamaged leaves could attract S. erinacea. Our research suggests that H. punctata can attract insect predators by releasing herbivory-induced volatile organic compounds, and for the first time we found ferns may also have indirect defense mechanisms using volatile organic compounds.

To answer these questions, the volatiles from healthy and wounded H. punctata (damaged by larvae of noctuid) were compared using GC-MS and determined the electroantennogram (EAG) responses of antennae of S. erinacea to different standards of volatile compounds identified from the GC-MS analysis. The behavior of male and female adults of S. erinacea to these volatile compounds was also investigated using a Y-tube olfactometer. Our findings provide insights into the mechanism underlying the attraction of S. erinacea to H. punctata.

Field Survey of Visitation
A field survey of the visitation and the behavior of S. erinacea had been done in an H. punctata population, in a valley in southern Hunan province, the H. punctata individuals were classified judging by the appearance of the leave into two groups: 41 healthy individuals (HI) and 34 damaged individuals (DI, meaning damaged by phytophagous insects). The S. erinacea usually stays on H. punctata for hours to a whole day during one visitation, thus the number of S. erinacea on each H. punctata individual was recorded in the morning. The observation lasted for seven days.

Collecting and Analyzing Volatile Organic Compounds from Hypolepis punctata
VOCs were sampled in naturally growing Hypolepis punctata populations in Fengluanxi (110 • 7 36 N, 29 • 19 50 E), Hunan Province. Several individuals of H. punctata with similar growth status and size that showed no sign of insect damage were selected, and the larvae of Bertula hadenalis, a species of noctuid were selected as the herbivorous insect. We divided all H. punctata individuals into three groups: healthy individuals (HI), undamaged individuals that were growing close to damaged individuals (UICD), and damaged individuals (DI, meaning damaged by B. hadenalis). Prior to volatile sampling, one leaf from each individual of Hypolepis punctata was safely enclosed in a fine mesh nylon bag. For the DI group, we added two to three two-instar larvae of Bertula hadenalis into each bag, and for the HI group, no larvae were placed into the bags. Additionally, the HI group individuals were more than 10 m away from the plants included in the DI group. The individuals of the UICD group were similar to the HI group, as no larvae were added to the bags, but the distance between the UICD individuals and the DI individuals was less than two meters. The mesh bags remained affixed to the leaves for 24 h before volatile sampling. During VOC sampling, the mesh bags and larvae were removed, and each leaf was enclosed in a Nalophan bag (Toppits, Minden, German) with openings at both ends. One end was fastened securely to the fern stipe together with a glass tube (diameter 4 mm, length 15 cm) filled with activated charcoal, and the other end was connected to another glass tube (diameter 4 mm, length 10 cm) containing 30 mg adsorbent (Porapak Q). The end of the adsorbent tube was connected to an airflow meter (model LZB-3W, 100-1000 mL/min, Shanghai Hange Biosciences Corporation, Shanghai), which was connected to an air pump (model QC-1S, Beijing Municipal Institute of Labor Protection, Beijing) set to a flow rate of 400 mL/min for four hours. Each part mentioned was connected using laboratory-grade PVC tubing. In the negative control experiments, the Nalophan bags were empty. In each location, there were four repeats of the experimental groups (DI, HI, and UICD) and two of the negative control group, and tests were conducted concurrently from 9:00 am to 2:00 pm. Compounds were eluted from the adsorbent tube with 600 µL of n-hexane and stored at −20 • C until analyzed. 0.4 µL of the eluent was analyzed using a Shimadzu GC-MS (GCMS-QP2010) with a RESTEK Rtx ® -5MS column (30 m × 200 mm, film thickness 0.25 mm). The carrier gas used was helium with a flow rate of 2.0 mL/min. The GC-MS was programmed as follows: injector 250 • C, initial column temperature held at 50 • C for 3 min, increased at 10 • C min −1 to 210 • C, maintained for 1 min, increased at 20 • C min −1 to 250 • C, and held for 3 min.
We used mass spectra from MS libraries (NIST05, NIST05s, and WILEY7) for tentative peak identification, and only those compounds with a higher concentration than in the control samples were considered. Retention indexes, which were used to convert retention times into system-independent constants, were taken from studies that used the Rtx ® -5MS column, as well as those that used the similar HP-5MS column. We compared the composition of VOCs in the individuals by calculating proportions of the contribution of identified compounds, which in each individual were calculated by dividing the area of each peak by the total area of all the peaks, except the peaks of contaminants, and multiplying by 100.

Electroantennogram Recordings
The responses of antennae of both male and female Sclomina erinacea to certain volatile compounds of Hypolepis punctata were tested using electroantennogram recordings (EAG). Nymphs of Sclomina erinacea were collected from Fengluanxi (110 • 7 36 N, 29 • 19 50 E), Hunan Province, and grown to adults under laboratory conditions fed with the larvae of Tenebrio molitor (25 • C, 60% RH). The compounds used for testing (five to be herbivory induced compounds and two green leaf volatiles) were selected according to the results obtained in the VOCs analysis, all the compounds used in this experiment are listed in Table 1. In this experiment, an entire antenna was carefully removed from the head of living adult Sclomina erinacea; 4/5 the length of the scape and 1/10 the length of the flagellum was also removed, and both sides of the antenna were attached to electrodes covered with electrode gel (Spectra 360, Parker Labs, Newark, NJ, USA). Ten µL of hexane (as control) or a hexane solution composed of one of the tested seven compounds (10 µg/µL, 100 µg) was dotted on a piece of filter paper (5 mm × 40 mm), which was then inserted into a glass Pasteur pipette. The tip of the pipette was inserted about 4 mm into a small hole in the wall of a glass tube, into which the antenna with the electrodes was inserted 3 cm from the open end of the tube without touching the inner wall. A constant airflow (1 L/min), passing through the glass tube and over the antenna, was generated and controlled by a Syntech CS-55 air stimulus controller, which also controlled the airflow through the pipette. During scent stimulation, an airflow of 40 mL/min was pumped through the Pasteur pipette into the glass tube containing an antenna for 1 s. Each antenna was stimulated two or three times with each compound individually at 60-s intervals, and each compound was repeated using eight different antennae from each sex (male and female). Hexanal was used as the reference standard, which we applied to stimulate the antennae at the beginning and end of each stimulation to correct for the loss of antennal sensitivity. For the correction, we assumed that the decrease of the antennal sensitivity had a linear relationship with time; therefore, data were normalized as follows: where rEAG stands for relative EAG response; EAG(A) stands for the amplitude (mV) of the EAG response to compound A; EAG(std1) stands for the EAG response to the reference standard at the beginning of each stimulation; EAG(std2) stands for the EAG response to the reference standard at the end of each stimulation; RT(A) stands for the time when the stimulation was done with compound A; RT(std1) stands for the start time of the stimulation using the reference standard at the beginning of each stimulation; RT(std2) stands for the end time of the stimulation using the reference standard.

Y-Olfactometer Experiment
Olfactory behavior of Sclomina erinacea towards seven different volatile organic compounds (Table 1) produced by Hypolepis punctata was tested using a glass Y-tube olfactometer, and an additional file shows a sample graph of the whole device (Supplementary Materials, Figure S1). Two pear-shaped glass bottles where the odor compounds were placed were connected to each end of the arms of the Y-tube (ID 25 mm; stem 6 cm; arms 15 cm at a 142 • angle to the stem). Both the Y-tube and the pear-shaped glass bottles were put inside a top opening square box (length 50 cm; width 35 cm; height 12 cm), made of paper and painted white to reduce environmental disturbance. The contents of the box were illuminated by three 30 W fluorescent lamps. Air cleaned by active charcoal and humidified by passing through distilled water was pumped through each pear-shaped bottle into each arm at a rate of 400 mL/min using an air pump (model QC-1S, Beijing Municipal Institute of Labor Protection, Beijing). Cuboid agar blocks (2%; 1.5 cm length; 1.5 cm width; 0.5 cm height) placed in the glass bottles were used to carry each compound and release them into the Y-tube. Each compound was diluted in hexane. For each test, 5 µL of the solution (10 µg/µL, 50 µg) was injected into one agar block, which was then placed into one of the two pear-shaped glass bottles. Another pear-shaped glass bottle holding an agar block containing 5 µL of hexane was used as the control group. Then, one adult insect was released from the opening of the stem into the Y-tube. Choices were scored when adult insects went 2 cm from the middle of the Y-tube into one arm, and an additional picture shows the position of an adult when it just reached the score point (Supplementary Materials, Figure S2). If an insect did not make any choice after 10 min, then the test was repeated using another insect individual until a choice was finally made. Each compound was tested using at least seven adults of each sex, and each test was repeated three times. To reduce the disturbance of uncertain factors, the Y-tube was cleaned once every five tests using absolute ethanol, and the position of the experimental group and the control group was also exchanged.

Data Analysis
Data of the field study, VOCs proportions, EAG responses, and the Y-tube tests were analyzed using software SPSS 19.0. G-test was used to analyze the average relative VOCs amounts, Duncan's multiple range test was used to compare EAG responses triggered by different compounds and G-test was used to compare the selection of different groups in the field study and the Y-tube test.

The Visitation Observation
The visitation preference difference of S. erinacea individuals between the healthy and the damaged H. punctata in the field observation is shown in Table 2, which demonstrate that, during the seven days of field observations, S. erinacea individuals had a strong preference for the damaged Hypolepis punctata (p < 0.05) by herbivory for five days in the total seven days, indicating that the herbivory of phytophagous insects on H. punctata can affect the behavior of the assassin bug S. erinacea. The p Values were generated using G-test for the comparison of the Sclomina erinacea individuals' visitation preference difference between the healthy and the damaged Hypolepis punctata, of which the individual numbers were 34 and 41 respectively. Table 3 presents the relative percentages of each volatile organic compound collected from healthy individuals (HI), undamaged individuals that are close to damaged individuals (UICD), and damaged individuals (DI, meaning damaged by Bertula hadenalis) of Hypolepis punctata. Twenty-five VOCs were detected in total, with 15 from the HI group, 12 from the UICD group, and 11 from the DI group. The compounds shared among all three groups were α-pinene, 1-octen-3-ol, and nonanal. Six terpenes were detected: αpinene, β-myrcene, sabinene, α-farnesene, zingiberene, and linalool. Figure 1 illustrates the comparison of relative percentages of nine VOCs from damaged and undamaged Hypolepis punctata. The relative percentages (divided by the total area of all the peaks) of five of the nine VOCs released from damaged individuals were greater than those released from healthy individuals, significantly for three of them (β-myrcene, linalool, α-pinene, and nonanal, p < 0.05), marginally significantly for α-pinene (p = 0.0565), and insignificantly for 1-octen-3-o (p > 0.05). Thus, these five compounds (β-myrcene, linalool, α-pinene, nonanal, and 1-octen-3-ol) were assumed to be herbivory-induced volatiles and were selected for the following test in this study.

VOCs of Herbivore Damaged and Undamaged Hypolepis punctata
The other four compounds (hexanal, trans-2-hexenal, 3-hexen-1-ol, and 1-hexanol) are green leaf volatiles (GLVs), and the relative amounts of these four compounds present the form opposite to that of the assumed herbivory induced volatiles. Figure 2 shows the relative percentages of these four GLVs sampled from individuals of the HI, UICD, and DI groups. The individuals of the HI group emitted the most substantial amount of these GLVs, which were released in smaller amounts by individuals of the UICD group, and were the lowest emitted from individuals of the DI group. This indicates that after H. punctata were damaged by B. hadenalis larvae, the release of some GLVs was reduced. Figure 3 illustrates the relative EAG responses of the antennae taken from Sclomina erinacea to tested compounds. The antennae of both male and female S. erinacea shows the highest responses to trans-2-hexenal compared with other tested chemicals (p < 0.05). The elicited relative EAG responses of the male antennae in descending order are trans-2-hexenal, 1-octen-3-ol, nonanal, linalool, hexanal, β-myrcene, α-pinene, and hexane. For antennae removed from female insects, the EAG responses in descending order are trans-2-hexenal, nonanal, hexanal, 1-octen-3-ol, linalool, β-myrcene, hexane, and α-pinene. All chemicals except for α-pinene and β-myrcene elicit significantly higher relative responses in both male and female antennae than hexane (p < 0.05). Linalool, α-pinene, trans-2-hexenal, and 1-octen-3-ol elicit higher relative responses in males than in females, and response to 1-octen-3-ol is significantly different (p < 0.05). The other three compounds elicit responses that are higher in females than in males (p > 0.05).    The other four compounds (hexanal, trans-2-hexenal, 3-hexen-1-ol, and 1-hexanol) are green leaf volatiles (GLVs), and the relative amounts of these four compounds present the form opposite to that of the assumed herbivory induced volatiles. Figure 2 shows the relative percentages of these four GLVs sampled from individuals of the HI, UICD, and DI groups. The individuals of the HI group emitted the most substantial amount of these GLVs, which were released in smaller amounts by individuals of the UICD group, and were the lowest emitted from individuals of the DI group. This indicates that after H. punctata were damaged by B. hadenalis larvae, the release of some GLVs was reduced.  Figure 3 illustrates the relative EAG responses of the antennae taken from Sclomina erinacea to tested compounds. The antennae of both male and female S. erinacea shows the highest responses to trans-2-hexenal compared with other tested chemicals (p < 0.05). The elicited relative EAG responses of the male antennae in descending order are trans-2hexenal, 1-octen-3-ol, nonanal, linalool, hexanal, β-myrcene, α-pinene, and hexane. For antennae removed from female insects, the EAG responses in descending order are trans-  12,978 igure 3. Relative EAG responses of Sclomina erinacea antenna to eight compounds. The Y-axis is the average relative EA esponse of the antennae of Sclomina erinacea to eight selected compounds, which are represented by numbers 1-8 on t -axis: 1 = hexane, 2 = α-pinene, 3 = trans-2-hexenal, 4 = hexanal, 5 = linalool, 6 = β-myrcene, 7 = nonanal, 8 = 1-octen-3ach insect sex is represented with a unique color, and the data was analyzed using Duncan's multiple range test with ach sex between compounds. The uppercase or lowercase letters on top of each column are placed for the comparison e EAG responses of males or females to each chemical, the uppercase letters are for males and the lowercase letters r females. Columns with one or more common letters on top represent there is no significance (p > 0.05) between t AG responses of antennae of one sex to each corresponding chemicals , otherwise the EAG responses to ea orresponding chemicals for each column(s) with unique letter(s) is(are) significantly different from other chemicals. T ymbol * means there is significant difference between the EAG responses of males and females to a chemical.

Y-Olfactometer Experiment
The number of choices of Sclomina erinacea among the seven tested comp versus hexane using a Y-olfactometer was shown in Table 4, while the percent of is shown in Figure 4. To male S. erinacea, nonanal, β-myrcene, and trans-2-hexenal s significant attractiveness (p < 0.05, Table 4). As to females, only linalool showed sig attraction (p < 0.05, Table 4). However, although the difference was not significant, f prefer trans-2-hexenal (18 vs. 9 times, Table 4), and males prefer linalool (16 vs. 8 Table 4) than towards hexane. Each insect sex is represented with a unique color, and the data was analyzed using Duncan's multiple range test within each sex between compounds. The uppercase or lowercase letters on top of each column are placed for the comparison of the EAG responses of males or females to each chemical, the uppercase letters are for males and the lowercase letters are for females. Columns with one or more common letters on top represent there is no significance (p > 0.05) between the EAG responses of antennae of one sex to each corresponding chemicals, otherwise the EAG responses to each corresponding chemicals for each column(s) with unique letter(s) is(are) significantly different from other chemicals. The symbol * means there is significant difference between the EAG responses of males and females to a chemical.

Y-Olfactometer Experiment
The number of choices of Sclomina erinacea among the seven tested compounds versus hexane using a Y-olfactometer was shown in Table 4, while the percent of choices is shown in Figure 4. To male S. erinacea, nonanal, β-myrcene, and trans-2-hexenal showed significant attractiveness (p < 0.05, Table 4). As to females, only linalool showed significant attraction (p < 0.05, Table 4). However, although the difference was not significant, females prefer trans-2-hexenal (18 vs. 9 times, Table 4), and males prefer linalool (16 vs. 8 times, Table 4) than towards hexane. A G-test was used to compare the selection of different groups, and * indicates that there is a significant difference (p < 0.05) between the choices of the control group and the experimental group.

Discussion
Our study provides the first clear evidence of indirect defense mechanisms in ferns. Once wounded by herbivorous insects, H. punctata individuals were significantly more attractive to S. erinacea compared with healthy individuals in the field (Table 2). Furthermore, the damaged individuals produced a significant amount of novel volatiles compounds, which were not detected in healthy individuals of this fern species. Electroantennogram measurements and behavioral experiments using volatile standards indicated that these novel volatiles compounds attract S. erinacea.

Volatiles Induced from H. Punctata
Our results illustrated that there was significantly more β-myrcene, linalool, nonanal, and marginally significantly more α-pinene secreted from wounded H. punctata in comparison with intact individuals. The average amount of secreted 1-octene-3-ol was not significantly different between the HI and DI groups; however, the average amount of 1-octene-3-ol was higher from wounded than from healthy plants. Linalool [23], β-myrcene [24], α-pinene [25], nonanal [26], and 1-octene-3-ol [27] have been shown to impact insect behavior. Among them, α-pinene, linalool, and 1-octen-3-ol have been found to be able to attract predators for plant defense [9,28,29]. Therefore, combined the phenomenon with the result of our field observation, we believe that these induced secondary metabolites are volatiles produced in the fern, and they may play an essential role in the indirect defense of H. punctata by attracting natural enemies.
It is metabolically expensive for plants to synthesize induced volatiles, and sometimes the synthesis would result in a reduced or halted production of other common volatiles [21]. In this study, the four green leaf volatiles were detected in only healthy individuals but were not detected in or only detected in small amounts in wounded H. punctata or nearby healthy individuals. We speculate that to synthesize induced volatiles, H. punctata reduced the release of green leaf volatiles (Figure 4). This is a metabolic trade-off in the plant.
Plants can communicate via volatile chemicals. Studies have shown that when plants are injured, adjacent plants initiate the corresponding mechanism of chemical defense [10][11][12]. In this study, β-myrcene and linalool were found in healthy H. punctata individuals near wounded individuals. Wounded H. punctata likely communicated the wounding chemically with other individuals to induce them to produce volatiles. However, this hypothesis needs further exploration.

Electrophysiological Experiments Using Volatile Standards
Electrophysiological experiments were used to accurately determine whether the insects sense chemicals through smell, which is an excellent method to study the relationship between insects and volatiles. In particular, this allowed us to determine which compounds in the tested volatiles are likely to affect insect behavior [30]. Among the selected volatile standards, five were induced chemicals, and two were GLVs from healthy leaves. All volatiles except α-pinene and β-myrcene generated stronger electrophysiological signals than the control, n-hexane, in the antennae of S. erinacea, suggesting that S. erinacea may at least sense some of the induced substances produced in the fern and be attracted to assist the fern (the indirect defense mechanism). It is possible that not all volatiles produced by H. punctata can be sensed, which might be why two of the identified compounds did not significantly stimulate the antennae.
Additionally, there might be other unknown functions for these volatiles. Studies have shown that α-pinene is a secondary metabolite with direct defense activity, which has been shown to reduce the number of larvae and adults Tribolium castaneum and decrease their food consumption and food utilization rate [31]. To our knowledge, there has been no report on the role of β-myrcene for effective defense in plants. In addition, it is possible that some of the induced components in our study may play a role in the chemical communication between plants of the same species because some of these induced chemical components were detected in individuals near the wounded plants. Green leaf volatiles (GLVs) are a kind of "common smell" emitted by many unwounded plants [32,33]. The releasing of Green leaf volatiles was not insect-wounding-needed. In our study, however, some of the GLVs (e.g., trans-2-hexenal) were able to generate electrophysiological signals in the antennae, and, notably, adult male and female S. erinacea were significantly more sensitive to trans-2-hexenal in EAG measurements (p < 0.05) than other compounds. This suggests that even before wounding, the insect may be able to sense the presence of H. punctata or plants with similar volatiles. This implies that S. erinacea may have adapted to sense non-induced plant scents to locate its habitat or proliferation site near potential hunting grounds, and then precisely locate its prey based on the induced chemical compounds, which requires further study.
In addition, there were no significant differences in the EAG measurements of other compounds between male and female S. erinacea, except for with 1-octen-3-ol, suggesting that the induced volatiles may not affect the reproductive behavior of S. erinacea, and the leading role of the volatiles is to attract the insects to catch pests.

Behavioral Experiments Using Volatile Standards
Compared with electrophysiological experiments, behavioral experiments can be easily influenced by environmental factors, especially when the tested compounds are unstable; the results are also influenced by the health of the tested insects. Furthermore, it is highly challenging to simulate the natural conditions of the olfactory condition entirely. Even so, among the seven tested volatiles, four had significantly stronger attractive activity on one sex of S. erinacea compared with the control, indicating that the induced volatile compounds attract S. erinacea with more differential activity observed between sexes, especially when compared with the EAG experiments, which is also similarly reported in one other related study [34]. Further investigations are needed to have a better understanding of the behavioral patterns between male and female S. erinacea individuals.
Among the volatile compounds that attract S. erinacea, nonanal and 2-hexenal were from undamaged green leaves, and β-myrcene and linalool were induced after wounding. The attractive activity by the remaining three compounds (excluding β-myrcene) is in line with the results of the electrophysiological experiments, suggesting that the attractiveness of H. punctata to S. erinacea may be through an indirect defense mechanism. The fact that green leaf volatiles can attract S. erinacea also implies that the insect may establish its habitat based on constitutional (not induced) smells from H. punctata. However, βmyrcene, which had a similar EAG value to hexane, was found to be attractive to adult male S. erinacea. This may be due to the reason that fewer odor sensilla that are sensitive to β-myrcene are in the antennae of the adult males than those sensilla sensitive to other compounds. Nonetheless, these sensilla still generate signals that impact the behavior of male S. erinacea. More studies are needed to elucidate the specific reasons for the attraction of males to β-myrcene. Our result also suggested that the attractiveness of these volatiles to S. erinacea seems to be irrelevant with their relative amounts (Table 4).
Our findings explain our preliminary field observations where we observed that noctuid larvae were heavily preyed on by S. erinacea when eating H. punctata. However, the effect of this defense mechanism is also dependent on the severity of the pest damage and the population density of S. erinacea. Therefore, further studies are needed to investigate the quantitative and dynamic relationship between noctuid larvae and S. erinacea.

Conclusions
Our data provide direct evidence that the herbivory-damaged H. punctata has more attractiveness to S. erinacea than the healthy individuals in the field, and the volatiles from H. punctata can attract S. erinacea, especially the induced compounds. This supports the proposition that ferns may also employ indirect defense mechanisms using volatile organic compounds. This mechanism is different from ants being attracted to fern species using nectar as a consumable reward [17] because H. punctata does not provide a direct reward to S. erinacea, yet still creates a mutually beneficial relationship with S. erinacea. Thus, we discovered an indirect defense mechanism of a fern species, and whether the volatile attraction mechanism found in this study resulted from long-term co-evolution still needs further research.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/insects12110978/s1, Figure S1: A sample graph of the whole device of the Y-olfactometer, Figure S2:

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The datasets used during the current study are available from the corresponding author upon request.