Radioprotective Effects on Late Third-Instar Bactrocera dorsalis (Diptera: Tephritidae) Larvae in Low-Oxygen Atmospheres

Simple Summary The oriental fruit fly Bactrocera dorsalis Hendel is a highly invasive fruit fly that causes extensive damage to many fruits and vegetables. Irradiation treatment is an economically effective and promising treatment measure. However, treatment efficacy is affected by the presence of low oxygen, i.e., mangoes are treated in modified atmosphere package. In order to investigate the reduced (radioprotective) effects on insects and determine its critical O2 level, third-instar B. dorsalis larvae were irradiated by X-rays at the doses of 8 to 64 Gy with intervals of 8 Gy. The treatments were conducted under ambient air or low-oxygen atomospheres (0%, 2%, 4%, 6%, 8% O2 and nitrogen). No adult emergence from treatments at 64 Gy in pure nitrogen or 56 Gy under other atmospheres, resulted in significant difference in tolerance. The results from statistical analyses indicate that differences in tolerance to radiation were significant in 0% and 2% O2 but insignificant in 4%, 6%, and 8% O2 environments when compared with radiation in ambient air. Therefore, the critical threshold is an O2 level of ≥4% and <6%, but a maximum radiation dose of 14 Gy can compensate for the radioprotective effects when the oriental fruit fly is treated in low-oxygen atmospheres. Abstract Ionizing radiation creates free radicals, the effect of which is enhanced by the presence of oxygen; a low oxygen level produces radioprotective effects for insects compared with irradiation in ambient air. Modified (controlled) atmosphere packaging is used for maintaining quality and shelf-life extension; therefore, treatment efficacy may be affected, and there is a need to determine the critical O2 levels that may cause radioprotective effects. Late third-instar Bactrocera dorsalis (Hendel) larvae were irradiated in bags filled with ambient or low-oxygen air (0%, 2%, 4%, 6%, 8% O2) and were exposed to radiation doses of 8 to 64 Gy with intervals of 8 Gy. Efficacy was measured by the prevention of adult emergence. Dose–response data on mortality (failure of adult emergence) were analyzed via two-way ANOVA (analysis of variance), ANCOVA (analysis of covariance), and probit regression. The difference in radiotolerance was only significant in 0% O2 atmospheres through two-way ANOVA; therefore, the 95% confidence limits (CLs) of lethal dose ratios at LD99 were used to determine significant differences between treatments at different O2 levels. The differences in radiotolerance were significant in 0% and 2% O2 but insignificant in 4%, 6%, and 8% O2 environments when compared with radiation in ambient air. The critical threshold of radioprotective effects for late third-instar B. dorsalis larvae is an O2 level of ≥4% and <6%, but a maximum radiation dose of 14 Gy can compensate for this effect during phytosanitary irradiation treatment.


Introduction
The oriental fruit fly, Bactrocera dorsalis Hendel (Diptera: Tephritidae), is a highly invasive species with over 300 known commercial/edible and wild hosts. It is currently found in at least 65 countries, including most of Asia, much of the sub-Saharan countries of Africa, parts of America, Oceania, and Europe, but it is of quarantine significance to many other countries [1][2][3]. In China, the oriental fruit fly is one of the most destructive quarantine insects of tropical and subtropical fruits and vegetables, causing severe losses to citrus, guava, carambola, and mango [4,5]. Infested commodities are normally required to undergo phytosanitary treatments before export to regulated or quarantine areas [6].
Phytosanitary measures, including fumigation, temperature, and irradiation treatments, are currently used for shipped commodities, and the use of phytosanitary irradiation (PI) treatment has increased in recent years due to its advantages over other treatments; as a result, the global volume of different fresh products was approximately 30,000 tons in 2017 [3,4,[7][8][9]. In order to develop treatment schedules and facilitate the application of PI treatment, many kinds of tephritid fruit fly third-instar larvae have been used for conducting radiotolerance and confirmatory studies, such as the oriental fruit fly, B. dorsalis [4,[10][11][12]; the Mediterranean fruit fly, Ceratitis capitata Wiedemann [10,12]; the melon fly, Zeugodacus cucurbitae Coquillett [10,13]; and Z. tau Walker [14]. Currently, the International Plant Protection Convention (IPPC) is discussing draft standards related to irradiation treatment for B. dorsalis, Z. tau, and the genus Anastrepha to formulate international standards, as an annex to International Standard for Phytosanitary Measures (ISPM) No. 28 [15].
Fresh commodities are always allowed to be irradiated with insect-proof packing, so as to protect them from infestation, re-infestation, or contamination afterwards. Modified (controlled) atmosphere packaging (MAP) is widely used to improve the shelf life of fresh fruits by decreasing the O 2 level (below 8%) and/or elevating the CO 2 level (above 10%) [16][17][18]. Treatment under an O 2 level lower than that in MAP is used as a single or combined phytosanitary treatment measure, and a draft ISPM (Draft ISPM 2014-006: Requirements for the use of modified atmosphere treatments as phytosanitary measures) is under discussion for adoption [15,19,20]. Ionizing radiation creates free radicals, the effect of which is enhanced by the presence of oxygen; thus, high oxygen tension can enhance the effects of radiation. Conversely, a low oxygen level produces radioprotective effects for insects, so radiation in low-oxygen environments is considered to be the most significant factor affecting the efficacy of PI treatment, aside from the radiation dose itself [21,22].
Although reduced effects of radiation have been validated by several studies that conducted radiation treatment in <3-5% O 2 atmospheres, most of them performed comparisons between ambient air and pure nitrogen or MAP, whereas the difference in radiation effects was insignificant [23]. In addition, the PI treatment efficacy was not reduced when fruit flies were treated under hypoxia [12]. Therefore, more research is required to determine the critical O 2 levels that may cause radioprotective effects for insects and at which point the efficacy of the recommended doses for commercial PI treatment may decrease when conducting treatments in low-oxygen environments [21][22][23]. Thus, the objective of this research was to investigate the critical O 2 level (threshold value) that may induce radioprotective effects for late third-instar B. dorsalis larvae and to estimate the radiation dose that compensates for this reduced efficacy.

Insect Rearing
The insects used in this research were originally collected in a mango orchard in Chongzuo city, Guangxi Zhuang Autonomous Region, China in July 2018, and then reared in the Laboratory of Phytosanitary Treatment and Equipment, Chinese Academy of Inspection and Quarantine in Beijing, China. Late third-instar B. dorsalis larvae that emerged from mango fruits were collected and transferred to plastic boxes containing moist sterile sand for pupation. About 10 days later, the pupae were placed in the rearing cages (40 × 40 × 50 cm) for adult emergence; the adults were fed with sterile water, fresh fruit pulp, and a solid mixture of sucrose and hydrolyzed yeast (3:1). Eggs were collected by placing papaya peel domes (emptied of fruit pulp) in the cages, and larvae were reared on artificial diets, the components of which were described by Liu et al. [24]. The rearing room was controlled at a temperature of 24-26 • C with 50-70% relative humidity and a photoperiod of 14:10 h (L/D). The emerged late third-instar larvae (about 8 days old) in the third to sixth generations were used for conducting all the radiation treatments.

Handling of Larvae
Two-liter gastight air bags, purchased from Dalian Delin Gas Packaging Co., Ltd. Dalian, China, were used for conducting radiation treatments in hypoxia or ambient air. For each treatment, about 100 late third-instar larvae were wrapped in a small net bag and placed into an air bag through its opening, followed by sealing of the bag, exhausting all the air with a one-liter air-tight syringe, and injecting pure nitrogen (<0.001% O 2 ) or controlled atmosphere (CA) (2%, 4%, 6%, or 8% O 2 and nitrogen) (Beijing Green Oxygen Tiangang Technology Development Co., Ltd. Beijing, China) into the bag and holding it for one minute. The procedure was repeated three times to purify the gas in the air bag; the bags containing larvae and a volume of 0.7-0.8 L CA were subjected to irradiation treatment 15 min later. For irradiation treatment in ambient air, the larvae were wrapped and placed in the air bags without any additional procedure.

Radiation Treatments
All the radiation treatments were conducted in an RS-2000 Pro X-ray irradiator (Rad Source Technologies, Inc., Coral Springs, FL, USA) with operating parameters of 220 KV and 17.6 mA. In order to get a good dose uniformity ratio (1.05), the irradiator was equipped with a reflector placed in the bottom of the exposure chamber (width, 17 inch; depth, 15 inch; height, 17 inch) [25]. The bag containing larvae was placed in the exposure chamber to be irradiated at doses of 0 (control) to 64 Gy with intervals of 8 Gy. For each dose treatment, three bags (as replicates) were placed in the exposure chamber and irradiated at the same time. To measure and control the applied radiation doses, a RadCal dosimeter (model 2086, RadCal Corp., Monrovia, CA, USA) with a 10 × 6-0.6 Ion Chamber was placed near the insects. After each radiation treatment run, the cumulative dose was recorded from the dosimeter, and the irradiator was stopped when the cumulative dose was reached. The dose rate monitored in the study was about 5.0 Gy/min.
After treatment, all the bags were opened when the total sealing time reached about one hour; then, the larvae were removed and placed in sterile moist sand for pupation. Three weeks later, the numbers of late third-instar larvae, pupae, and adults were counted.

Data Analyses
Dose−response data on mortality (failure of adult emergence) were adjusted by using Abbott's formula [26] and then subjected to two-way ANOVA (analysis of variance) to analyze the individual effects of radiation dose and PO 2 (O 2 level: 0%, 2%, 4%, 6%, 8%, and 21%) and the interaction effects of dose × PO 2 ; means were compared by Tukey's multiple comparison tests [27]. Linear regression after analysis of covariance (ANCOVA) was also used to analyze the mortality data to compare the Insects 2020, 11, 526 4 of 10 radiotolerance of B. dorsalis treated in different O 2 atmospheres; all the data less than 100% and the lowest dose causing 100% mortality were used in the analyses [13,27]. All the data derived from treatments below 56 Gy and from the control were subjected to probit analyses using the PoloPlus 2.0 program to estimate the minimum dose for preventing adult emergence at each environmental O 2 level via probit and logit models (using the non-transformed dose) [28]. In addition, to compare the significance of radiotolerance between treatments at different O 2 levels, pairwise comparison tests were performed by calculating the 95% confidence limits (CLs) of the lethal dose ratios at LD 99 (the minimum lethal dose leading to 99% mortality at a specific confidence level, i.e., 95%) [29]. If the 95% CL excludes 1, then the LD 99 values are significantly different [28,30].

Effects of Radiation Dose and Oxygen Level
The percent mortality of late third-instar B. dorsalis larvae treated under low oxygen and ambient air were found to be significantly affected by radiation doses (F 7,143 = 1089.0, p < 0.0001), PO 2 (F 5,143 = 4.6, p = 0.0026), and dose × PO 2 interactions (F 35,143 = 1.7, p = 0.0255) when the dose−response data were analyzed by two-way ANOVA (Table 1). With increasing radiation dose, adult emergence declined across all PO 2 treatment groups, and no insects successfully emerged as adults when treated at 64 Gy in pure nitrogen (<0.001% O 2 ) or at 56 Gy in other environments (2%, 4%, 6%, 8%, and 21% O 2 ). For treatments in pure nitrogen, the mean mortality at all doses (65.0 ± 31.6%), as well as that at 56 Gy (95.5 ± 2.3%), was significantly lower than those under other treatments, suggesting that the radiation effect is decreased significantly under extreme hypoxia. Meanwhile, other statistical methods were needed to compare the significance in radiotolerance between treatments in ambient air and low-oxygen atmospheres. Table 1. Radiation effects on the prevention of adult emergence as a result of irradiating late third-instar Bactrocera dorsalis larvae with X-rays (0-64 Gy) under ambient air and low-oxygen atmospheres.

Linear Regression
The results obtained from ANCOVA and linear regression ( Table 2) showed that all the coefficient of determination (R 2 ) values were larger than 0.95 (maximum of 0.985), which implies that the regression fit the data well. As the interaction effect between dose and PO 2 was significant (F 5,114 = 2.77, p = 0.0212), the minimum dose for 100% mortality was predicted by liner regression to compare the relative radiotolerance among treatments in a low-oxygen atmosphere and in ambient air. The estimated dose leading to 100% mortality decreased gradually with the increasing O 2 level from 0% to 4%, and then remained essentially unchanged (Table 2, Figure 1), suggesting that an O 2 level of 4% is likely to be the critical threshold of radioprotective effects for late third-instar B. dorsalis larvae. Furthermore, compared with that in ambient air, the minimum radiation doses for 100% mortality increased from 57.6 (in ambient air) to 58.3 (1.2%), 60.3 (4.7%), and 65.8 Gy (14.2%) in 4%, 2%, and 0% O 2 atmospheres, respectively. is likely to be the critical threshold of radioprotective effects for late third-instar B. dorsalis larvae. Furthermore, compared with that in ambient air, the minimum radiation doses for 100% mortality increased from 57.6 (in ambient air) to 58.3 (1.2%), 60.3 (4.7%), and 65.8 Gy (14.2%) in 4%, 2%, and 0% O2 atmospheres, respectively.

Probit Analyses
Probit analyses were used to analyze the dose-mortality data of late third-instar B. dorsalis larvae. The estimated values of LD99 and LD99.9968 (the minimum lethal dose achieving a mortality of 99.9968% at a specific confidence level, i.e., 95%) analyzed by the probit and logit model were very close; thus, only the parameters derived from the probit model, such as slope, intercept, heterogeneity, estimated lethal doses, and their 95% CLs, are listed in Table 3. A small heterogeneity factor and a narrow LD99 confidence interval clearly showed that the analysis fit the data well. The slopes increased but the estimated values of LD99 and LD99.9968 decreased sharply with increasing the O2 level from 0% to 4% and were almost unchanged until 21% (Table 3, Figure 1). Similarly, the mortality curves obtained from PoloPlus software ( Figure 2) can be divided into three groups, in which curve B (2% O2) is above curve A (0% O2) but below the merged curve clusters C (4%, 6%, 8%, and 21% O2). Furthermore, the 95% CLs of the lethal dose ratios at LD99 were pairwise compared between treatments at all O2 levels; the results (Table 3) showed that the radiotolerance of B. dorsalis irradiated in the 0% and 2% O2 environments was significantly greater than that of others, but it was insignificant when irradiated in 4%, 6%, or 8% O2 or in ambient air. Similar tendencies were obtained by comparing the estimated values of LD99.9968, though they were extrapolative values (Table 3, Figure 1. The estimated radiation dose for 100% mortality, LD 99 , and LD 99.9968 to prevent adult emergence from late third-instar Bactrocera dorsalis larvae irradiated in ambient air and 0%, 2%, 4%, 6%, and 8% O 2 environments.

Probit Analyses
Probit analyses were used to analyze the dose-mortality data of late third-instar B. dorsalis larvae.
The estimated values of LD 99 and LD 99.9968 (the minimum lethal dose achieving a mortality of 99.9968% at a specific confidence level, i.e., 95%) analyzed by the probit and logit model were very close; thus, only the parameters derived from the probit model, such as slope, intercept, heterogeneity, estimated lethal doses, and their 95% CLs, are listed in Table 3. A small heterogeneity factor and a narrow LD 99 confidence interval clearly showed that the analysis fit the data well. The slopes increased but the estimated values of LD 99 and LD 99.9968 decreased sharply with increasing the O 2 level from 0% to 4% and were almost unchanged until 21% (Table 3, Figure 1). Similarly, the mortality curves obtained from PoloPlus software (Figure 2) can be divided into three groups, in which curve B (2% O 2 ) is above curve A (0% O 2 ) but below the merged curve clusters C (4%, 6%, 8%, and 21% O 2 ). Furthermore, the 95% CLs of the lethal dose ratios at LD 99 were pairwise compared between treatments at all O 2 levels; the results (Table 3) showed that the radiotolerance of B. dorsalis irradiated in the 0% and 2% O 2 environments was significantly greater than that of others, but it was insignificant when irradiated in 4%, 6%, or 8% O 2 or in ambient air. Similar tendencies were obtained by comparing the estimated values of LD 99.9968 , though they were extrapolative values (Table 3, Figure 1). Thus, the sequence of radiotolerance is suggested as follows: 0% > 2% > 4% ≈ 6% ≈ 8% ≈ 21% O 2 atmospheres.
Compared with that in ambient air, the LD 99.9968 value was increased by 13.9 (17.2%), 8.0 (10.0%), and 1.8 Gy (2.2%) when larvae were irradiated in 0%, 2%, and 4% O 2 atmospheres, respectively. Therefore, the additional radiation dose that might compensate for the radioprotective effects is less than 13.9 Gy (nominally 14 Gy) when the oriental fruit fly is treated in low-oxygen atmospheres. Compared with that in ambient air, the LD99.9968 value was increased by 13.9 (17.2%), 8.0 (10.0%), and 1.8 Gy (2.2%) when larvae were irradiated in 0%, 2%, and 4% O2 atmospheres, respectively. Therefore, the additional radiation dose that might compensate for the radioprotective effects is less than 13.9 Gy (nominally 14 Gy) when the oriental fruit fly is treated in low-oxygen atmospheres.

Discussion
Radiotolerance in insects develops with their age and developmental time, so the most developed stage is the most radiotolerant when a common measure of efficacy is used [21]. The late third instar has been defined as the most tolerant stage of B. dorsalis larvae, as well as of other tephritid fruit flies, and was selected as the target stage for conducting dose-response tests in our research [4,[10][11][12][13][14]. The third-instar larvae were irradiated in a series of O2 levels with the same dose to compare the relative tolerance between treatments, and the prevention of adult emergence was selected as the common efficacy criterions [4,9,11,31,32]. Radiation tolerance in insects is modified by the oxygen level and increases in low-oxygen atmospheres; in particular, insects irradiated in very low oxygen (<1% O2) have repeatedly been shown to have an increased radioprotective response [10,23]. In our study, an increased radiotolerance of B. dorsalis in low-oxygen environments (0%, 2%,

Discussion
Radiotolerance in insects develops with their age and developmental time, so the most developed stage is the most radiotolerant when a common measure of efficacy is used [21]. The late third instar has been defined as the most tolerant stage of B. dorsalis larvae, as well as of other tephritid fruit flies, and was selected as the target stage for conducting dose-response tests in our research [4,[10][11][12][13][14]. The third-instar larvae were irradiated in a series of O 2 levels with the same dose to compare the relative tolerance between treatments, and the prevention of adult emergence was selected as the common efficacy criterions [4,9,11,31,32]. Radiation tolerance in insects is modified by the oxygen level and increases in low-oxygen atmospheres; in particular, insects irradiated in very low oxygen (<1% O 2 ) have repeatedly been shown to have an increased radioprotective response [10,23]. In our study, an increased radiotolerance of B. dorsalis in low-oxygen environments (0%, 2%, and 4% O 2 ) was confirmed by comparing mortality rates using two-way ANOVA (Table 1), the estimated minimum doses for 100% mortality by linear regression after ANCOVA (Table 2, Figure 1), and LD 99 by probit analysis (Table 3, Figures 1 and 2). The results are concordant with those from other radiation treatments in low-oxygen atmospheres against the Caribbean fruit fly, Anastrepha suspensa Loew [33]; C. capitata [34]; Z. cucurbitae [13]; Drosophila suzukii Matsumura [35]; the Oriental fruit moth, Grapholita molesta Busck [36]; the European corn borer, Ostrinia nubilalis Hübner [37]; and the cabbage looper moth, Trichoplusia ni Hübner [22].
Compared to those for radiation in ambient air, the minimum dose values for 100% mortality, LD 99 , and LD 99.9968 estimated in pure nitrogen were increased by 8.2 (14.2%), 9.2 (16.2%), and 13.9 Gy (17.2%), respectively (Tables 2 and 3 (Table 3) is very close to that reported by ) Gy), indicating that the experiments should be replicated and statistically analyzed to ensure that data are verifiable and reproducible [11,38]. The estimated LD 99.9968 values can be used as the minimum dose for PI treatment if they are validated by large-scale confirmatory tests [32,38], and our LD 99.9968 data (80.6 (76.3-85.6) Gy) ( Table 3) are very close to the values (84.1 (73.6-99.3) Gy) that were estimated and validated by Zhao et al. [4]; thus, a maximum radiation dose of 14 Gy can compensate for radioprotective effects during the PI treatment of oriental fruit fly. This explains why no radioprotective effects were observed when a radiation dose of 116 Gy (31.9 to 35.4 Gy higher than the LD 99.9968 estimates) was applied to late third-instar B. dorsalis larvae under severe hypoxia (0.3 ± 0.02% O 2 , 21.6 ± 0.1% CO 2 ) [12].
Commercial PI treatment in MAP or controlled atmospheres is more complicated [39]. The minimum applied dose should be equal to or higher than the maximum dose in the confirmatory tests; therefore, it is much greater than the LD 99.9968 estimates, and radioprotective effects are not observed as a result [12,21,31]. Moreover, although the treatment efficacy is reduced by low-oxygen atmospheres, it is increased by the presence of CO 2 , low temperature, and the long duration of modified atmosphere treatment after radiation treatment [11,12,32,38].
In order to define the critical threshold of radioprotective effects, radiation treatments in a series of oxygen levels should be carried out and the dose-response data subjected to statistical analysis to determine the statistically significant differences in radiotolerance [22,23]. The probit model is commonly used to analyze dose-response data [38,40]; however, the values of the slope obtained in this research (Table 3) are unequal, which means that the regression lines are unparallel, whereas they were parallel when Srimartpirom et al. treated B. dorsalis in MAP [11]. Therefore, the 95% CLs of the lethal dose ratios at LD 99 (ratio test) were used for comparing the relative radiotolerance [28,39]. As a result, oriental fruit fly irradiated in 0% and 2% O 2 showed significantly higher tolerance than that irradiated in ambient air and other low-oxygen atmospheres ( Table 3, Figures 1 and 2). There were no differences among 6% and 8% O 2 and ambient air; however, visible but insignificant tolerance differences were present between the treatments in 4% and 6% O 2 (Tables 2 and 3; Figures 1 and 2). To analyze the trends more intuitively, like how Chao et al. [23] predicted the critical O 2 level for the radiation treatment of T. ni, the estimated values (100% mortality, LD 99 , and LD 99.9968 ) at all O 2 levels in our testing were plotted in Figure 1. All three curves remained flat through the treatment at 21%, 8%, and 6% O 2 but turned up slowly for treatments at 6-4% O 2 and then steepened rapidly; therefore, the critical threshold of radioprotective effects is an O 2 level of ≥4% and <6% when late third-instar B. dorsalis larvae are irradiated in low-oxygen atmospheres. Since there is a small difference in LD 99 values of 1.1 Gy (1.95% of the radiation doses) between the treatments in 4% and 6% O 2 (Table 3), more precise research is required to reduce the range of critical values, for example, by conducting radiation testing in 4.5% or 5% O 2 atmospheres. Similarly, radioprotective effects have been observed in other fruit insects irradiated at <3-5% O 2 levels, such as A. suspensa [33], C. capitata [34], Z. cucurbitae [13], D. suzukii [35], G. molesta [36], O. nubilalis [37], and T. ni [22,23]. Furthermore, all these results support the APHIS' (Animal and Plant Health Inspection Service, United States Department of Agriculture) changes to administrative requirements, wherein the minimum O 2 level was reduced from 18% to 10% for conducting PI treatments [22,41]. However, the IPPC prohibits the use of PI treatment under low-oxygen atmospheres except for the treatment of G. molesta [20]; therefore, more insect species in different orders or families still need to be irradiated under low-oxygen environments to determine their critical threshold of radioprotective effects or the extra radiation dose that can compensate for the reduced efficacy. After that, a generic threshold of radioprotective effects can be established to facilitate the application of PI treatment of commodities with MAP or in low-oxygen environments.
Probit analysis has been widely used to analyze dose-response data for a number of fruit fly studies, and the ratios test is often used to determine the relative toxicity among a number of chemicals to determine the relative susceptibility of populations to pesticide resistance [4,11,14,29,[42][43][44]. In addition, confidence interval (CI) overlap, relative median potency, and one-way ANOVA (on LD 95 ) have been used to determine the significance of radiotolerance differences [11,23,30]. The 95% CLs of the lethal Insects 2020, 11, 526 8 of 10 dose ratios at a specific efficacy level, which are calculated automatically using computer software, are more general than the alternative statistics: Relative potency, which assumes that the regression lines are parallel; the CI overlap test, which should be used only when no alternative test exists; and one-way ANOVA, which requires at least three replicates of the LD X (for example, four replicates were conducted in radiation treatment of the cabbage looper moth, resulting in significant radioprotective effects determined in <0.1% and 2.5% O 2 atmospheres) [11,23,40]. The ratio test showed that the LD 99 value in 2% O 2 atmospheres is significantly larger than those in air and in 4%, 6%, and 8% atmospheres ( Table 3), but all the 95% CIs overlapped, suggesting that the ratio test is more sensitive than the CI overlap test, which provided a lower observed type I error rate of 0.004 to 0.005 [30]. Since probit-9 (LD 99.9968 at 95% CL), which is well known for use as the criterion for the PI treatment of insects, is an extrapolated value, the LD 99 that is calculated from the dose-response curve of tests in the most tolerant life stage is used as the minimum treatment level [38]; therefore, the 95% CLs of the lethal dose ratios at LD 99 are recommend for comparing the significance of tolerance in phytosanitary treatments.

Conclusions
In this study, radioprotective effects were demonstrated and their critical threshold was determined by irradiating late third-instar B. dorsalis larvae with 220 KV X-rays in ambient air and low-oxygen environments. The ratio test indicated that the differences in radiotolerance were significant in 0% and 2% O 2 but insignificant in 4%, 6%, and 8% O 2 environments when compared with treatment in ambient air; the critical threshold of radioprotective effects is, therefore, an O 2 level of ≥4% and <6%, but a maximum radiation dose of 14 Gy can compensate for this effect during PI treatment. We recommend that the 95% CLs of the lethal dose ratios at LD 99 be used for comparing the significance of tolerance in phytosanitary treatments, and a generic threshold of radioprotective effects should be established by testing a variety of insects.