Tsetse flies, Glossina pallidipes used in this study were obtained from the Biotechnology Research Institute (BioRI-KALRO). They were fed on alternating days on whole defibrinated bovine blood (collected from Farmers Choice Ltd., Nairobi, Kenya (S 01° 11.104′ E 036° 54.489′)). The blood was subjected to 1.0 kGy gamma irradiation (in a Gammacell 220 Excel irradiator (MDS Nordion, Ontario, Canada)) to ensure its sterility and then stored in 20 mL batches at −20°C (long-term storage). Before use, each frozen batch was thawed at 4°C then warmed to 37°C and presented to the flies (in an in vitro feeding method) through silicon rubber membranes (Langley, 1989).

The original G. pallidipes colony was derived from field collections between the Kenya-Uganda border near Tororo (Ciosi et al., 2014). The rearing was done at the Animal Rearing and Containment Unit (ARCU) at the International Centre of Insect Physiology and Ecology (icipe), under controlled conditions of 25 ± 2°C, 70% relative humidity (RH), and a photoperiod of 12 h light (L):12 h dark (D).

Glossina pallidipes pupae were maintained at 26.5°C until emergence. Six days following eclosion, two groups of males were selected: one control group (unirradiated) and the other group designated for SIT treatment (irradiated).

The six-day-old male flies that had been fed twice were irradiated in a Gammacell 220 Excel irradiator (MDS Nordion, Ontario, Canada). The central dose rate was calibrated using transfer alanine-electron spin resonance (ESR) dosimetry against a secondary standard with a 95% CI ± 2.7%. Dose distribution mapping was performed using Gafchromic HD810 film (95% CI ± 4%). After chilling at ≈4°C, the males were placed in a 110 × 50 mm cage for irradiation with 120 Gy in air (108–132 Gy) (Langley et al., 1974). The irradiated flies were exposed to 4°C for 6 h at RH 32–34.5% in a temperature-controlled incubator, and the unirradiated control group remained in the colony-holding conditions. Upon removal from the low-temperature chamber, the flies were immediately returned to colony-holding conditions in standard colony cages (18 cm x 7 cm x 4.5 cm) overnight for acclimatization before the bioassays. Any overnight mortality among the experimental flies in the colony cages was recorded. Thirty (n = 30) males of G. pallidipes were used per treatment and replicated three times for the bioassays.

For this study, 10 Metarhizium anisopliae strains (ICIPE 7, ICIPE 18, ICIPE 20, ICIPE 30, ICIPE 32, ICIPE 40, ICIPE 41, ICIPE 62, ICIPE 69, ICIPE 78) and one Beauveria bassiana strain ICIPE 603 were obtained from the International Centre for Insect Physiology and Ecology (icipe)’s germplasm maintained by the Arthropod Pathology Unit and were screened for their efficacy against tenerals of G. pallidipes. The isolates were sub-cultured on Sabouraud Dextrose Agar (SDA, Oxoid Ltd., Hampshire, United Kingdom) except B. bassiana, which was sub-cultured on Potato Dextrose Agar (PDA, Oxoid Ltd) and maintained at 25 ± 2°C in complete darkness until sporulation. Darkness is essential for the EPF’s consistent growth conditions (Pitt and Hocking, 2009) and enhanced sporulation (Inglis et al., 2001). The selected strains of M. anisopliae and B. bassiana do not produce resting spores in the traditional sense, such as zygospores in Zygomycota or chlamydospores in certain other fungi. Instead, they primarily form conidia (asexual spores), which are the main propagules used for infection and dissemination (St Leger, 1995). Conidia were harvested from three 3-week-old sporulated plates using sterile spatulas. The harvested conidia were mixed in 10 mL of sterile distilled water containing 0.05% (w/v) Triton X-100 (Sigma Aldrich, Inc., Missouri, United States) in a 30 mL universal bottle containing glass beads (5–10 beads of 3 mm in diameter per bottle) and vortexed for 5 min at about 700 rpm (using a ZX3 Advanced Vortex Mixer, VELP Scientifica Srl, Usmate, Italy) to produce homogenous conidial suspensions. Conidial concentrations were then determined using a new improved Neubauer chamber (Paul Marienfeld GmbH & Co. KG, Lauda-Königshofen, Germany) under the Leica DM2000 LED light microscope (Leica Microsystems (UK) Ltd). Prior to every bioassay, the conidia viability of each isolate was tested; where the conidial suspensions were then adjusted to a concentration of 3 × 10⁶ conidia mL⁻¹ then 0.1 mL of the suspension evenly spread on SDA or PDA plates. Plates were sealed with parafilm and incubated at 25 ± 2°C in complete darkness for 18 h. Each plate represented a replicate, and four replicates were done for each isolate and the treatments were arranged in a completely randomized design (CRD). After 18 h post-incubation, the plates were unsealed, fixed with a drop of fixative lactophenol cotton blue, and randomly covered with four sterile microscope coverslips (2 × 2 cm). Percentage conidial germination was determined from 100 randomly selected conidia on the surface covered by each coverslip under a Leica DM2000 LED light microscope (×40). Conidia were considered germinated when the length of the germ tube was at least twice the diameter of the conidium (Goettel and Inglis, 1997; Maniania, 1992). The germination rate determines whether the selected EPF isolate will be used for the bioassay. Over time, spore viability is affected by several factors; hence, it was monitored carefully. No EPF strain with ≤80% spore germination was used. If the spore germination rate fell below this threshold, the EPF strain was not used but was instead exposed to a susceptible host, then pin-isolated to regain its virulence, sub-cultured, and finally used for bioassays (Goettel and Inglis, 1997).

Due to the large quantities of fungi needed for the bioassays, mass production of the EPF isolates was done on rice substratum (separately for each isolate) in 600 mm × 350 mm Breathable Bio Control PP Bags (24 × 14) with double B filter (Unicorn Imp. & Mfg. Corp, Plano, Texas, US). The rice (2 Kg in each bag) was autoclaved for 1 h at 121°C and 15 psi, then allowed to cool in a Purifier Logic+ Class II, Type A2 Biosafety Cabinet (Labconco Corp, Kansas City, MO, United States), after which inoculation was done using 50 mL 3-day old culture of blastospores of the desired fungal isolate. The blastospores were prepared by aseptically scoping pure conidia into a 250 mL Erlenmeyer flask containing 50 mL of sterile liquid broth [30 gL⁻¹ glucose (Sigma Aldrich)], 30 gL⁻¹ yeast extract (Oxoid Ltd), 15 gL⁻¹ peptone (Oxoid Ltd) which was then incubated at 25 ± 2°C at 150 rpm for 3 days in a New Brunswick™ Innova® 44/44R incubator shaker (Eppendorf AG, Hamburg, Germany). The inoculated rice bags were incubated for 3 weeks at 25 ± 2°C and 40–70% relative humidity (RH), with continuous monitoring for quality control. After this, the rice with fungal growth was placed in basins (33 × 25 × 13 cm), to allow the conidia to dry at 25 ± 2°C, for 7 days. The conidia were harvested through agitating using MycoHarvester 6™ (VBS Ltd., United Kingdom; http://www.dropdata.net/mycoharvester/). The harvested pure conidia were packed in sterile zip-lock bags and then stored at 4 ± 2°C until used in downstream bioassays. Before any bioassays, conidia viability was assayed as described above.

Experimental insects (irradiated and unirradiated G. pallidipes) were infected with entomopathogenic fungi in separate batches using velvet carpet fabric layered with dry conidia that covered the inside of a cylindrical plastic tube (infection chamber) (95 × 48 mm) and had white nylon netting over one end (Dimbi et al., 2003; Wamiti et al., 2018). Dry conidia (0.1 g) were spread evenly onto the velvet material. Each batch consisted of 30 males (6 days old), which were enclosed within the infection chamber and exposed to the fungus for 12 h, after which they were transferred and placed into clean fungus-free cylindrical plastic tubes (95 × 48 mm), and maintained at 25 ± 2°C and 55% RH in a MIR-554 cooled incubator (PHC Europe B.V., Nijverheidsweg, Netherlands), at a photoperiod of 12 h Light: 12 h Dark. A total of 90 tenerals were eventually used (n = 30, replicated three times) for each treatment. The fungal_exposed flies were fed on bovine blood as described above. The control treatments (unirradiated control groups) were not exposed to any fungal strains and the experiments were run for 15 days, where the mortality of the flies was recorded. For all mortality observed, mycosis tests were conducted to confirm whether death was due to infection by the fungus used in the bioassays. The insect cadavers were surface_sterilized with 70% alcohol, rinsed three times in distilled water, and then kept separately in 90 mm Petri dishes lined with sterile moistened filter paper to allow for fungal outgrowth and verify if mortality could be attributed to the respective fungal isolates they were treated with. Mortality due to fungal infection was confirmed by the presence of hyphae and conidia on the surface of the cadavers. The median lethal time (LT₅₀) of the EPF isolates’ efficacy was calculated (15 days post fungi exposure) to further select the SIT-compatible EPF strains.

The colony diameter and radial growth rate of the best selected SIT-compatible EPF isolates (Metarhizium anisopliae isolates ICIPE 20, ICIPE 41, ICIPE 62 and ICIPE 78, and Beauveria bassiana ICIPE 603) were assayed under different temperatures, i.e., 15, 20, 25, 30, 35, and 40°C, respectively. To assess the effect of temperature on radial growth and colony diameter of these five selected candidates of EPF isolates, they were cultured on SDA plates except ICIPE 603 which was cultured on PDA and maintained at 25 ± 2°C under complete darkness as described above for 3 days to obtain mycelial mats. On the third day post-incubation, fresh SDA or PDA plates (9 cm diameter) were prepared. Using a marker pen, lines were drawn at the back of the plate to divide the plate into four quadrants as described by Agbessenou et al. (2021). Mycelial mats were cut from culture plates into round agar plugs using an eight-mm diameter cork borer. Each agar plug (ca. 5 mm thick) was then transferred onto the centre of a fresh SDA or PDA medium plate from which a similar size plug of media had been previously removed using the same cork borer. The plates with implanted mycelial plugs were sealed with Parafilm membrane and incubated in complete darkness at the defined temperatures, 15, 20, 25, 30, 35, and 40°C. Radial growth was recorded daily for 14 days using two cardinal diameters (a flexible ruler), through two orthogonal axes (the four quadrants) previously drawn on the bottom of each Petri dish to serve as a reference (Fargues et al., 1992). The experiment was replicated four times with each replicate originating from a different culture plate, where the radius of the growing fungi was measured, for all four quadrants. Fourteen days post-incubation conidia were then harvested by scraping the surface of the sporulated cultures from each plate using a sterile spatula. On the 14th day after taking the daily radius readings, the spore yield of each fungal plate was determined followed by germination assays of all the plates per temperature treatment. The harvested conidia were then suspended in 10 mL sterile distilled water containing 0.05% Triton X-100 and vortexed for 5 min at about 700 rpm (using a ZX3 Advanced Vortex Mixer, VELP Scientifica Srl, Usmate, Italy) to break conidial clumps and ensure a homogenous suspension. Conidial concentrations were quantified using an improved Neubauer hemocytometer under a light microscope as described above (Goettel and Inglis, 1997).

For conidia acquisition, retention and horizontal transfer bioassays, two potent M. anisopliae isolates ICIPE 20, and ICIPE 78 that were SIT-compatible/tolerant and responded well to the temperature from the above bioassay were used. Before evaluating whether EPF-exposed males and females could transmit conidia to their conspecific mates, we determined whether the amount of conidia carried by G. pallidipes individuals varied across time and between sexes. Hence, 0.1 g of dry conidia were evenly lined on a velvet carpet fabric in an infection chamber (95 × 48 mm cylindrical tube), then 10 G. pallidipes teneral males (conidia donors) were introduced into the infection chamber, were exposed to the fungus for 12 h, and then transferred into clean rearing cages with 10 healthy virgin fungus-free G. pallidipes females (conidia receivers). The conidia donors (males) were interacted with the conidia receivers (females) for 24 h then all the conidia donors were removed and kept separate (Dimbi et al., 2003; Wamiti et al., 2018). The conidia receivers were later collected at random (three female flies per day per replicate) on 1st, 3rd, and 5th day post exposure to fungus-challenged G. pallidipes males (conidia donors), then individually placed into universal bottles with 2 glass beads and 1 mL sterile distilled water containing 0.05% Triton X-100. The universal bottles with the EPF exposed flies were vortexed for 5 min at 700 rpm (using a ZX3 Advanced Vortex Mixer (VELP Scientifica Srl, Usmate, Italy) to release conidia from the insect’s body), then the conidia on their bodies quantified as already described above, using an improved Neubauer hemocytometer. This experiment was set in a completely randomized design, replicated three times and maintained in similar conditions as described above.

An inverse bioassay was set up whereby G. pallidipes females were used as conidia donors while the G. pallidipes teneral males were used as the conidia receivers. In this setup, 10 healthy virgin G. pallidipes females were contaminated with 0.1 g of dry conidia (as described above) and were later introduced into rearing cages with 10 fungus-free teneral G. pallidipes males (conidia receivers), after which the conidia donors (females) were removed after 24 h and the conidia receivers (males) collected at random as already described above. The conidia from the flies’ bodies were recovered and quantified as already described above. This experiment was also set in a completely randomized design and replicated three times. In all the bioassays, we considered fungus-free males and females as control treatments. Exposure of the flies to M. anisopliae ICIPE 20 and ICIPE 78 was done separately.

Another bioassay was set up as described above in the horizontal transmission where the insects were infected using the infection chamber. Two M. anisopliae isolates ICIPE 20 and ICIPE 78 were used in this bioassay. This experiment was done in two setups. In the first, teneral males were conidia donors and females as conidia receivers in one scenario, while in the second, females were conidia donors and males as conidia receivers (as described in section 2.7 above). Both the donors’ and the receivers’ treatment groups (for both scenarios) were observed daily for mortality for 15 days, and mycosis test was conducted on the recorded cadavers as described above. A similar setup of bioassay was done with male and female flies which were not exposed to any fungus and used as a control treatment. This experiment was set in a completely randomized design and replicated three times. Mortality due to fungal infection was confirmed through mycosis test as described above by assessing the presence of hyphae and conidia on the surface of the cadaver. The survival rates of both the males and females in the donor and receiver groups were then calculated.

The absolute radial growth (slope; mm/day) was determined by plotting the radial growth data against time (days) and fitted to a linear regression model (γ = βx + c) in Microsoft Excel (Microsoft 365), whereby γ represents the radial growth (indicating size of the mycelium at a given time), β represents the slope of the linear regression (indicates change of radial growth over time), x represents time (measured in days) and c represents the y-intercept of the linear regression line. The computed radial growth data was then fitted into both linear and nonlinear temperature-dependent models to establish the relationship between EPF growth rate and temperature. This association between the temperature regimes and the fungal growth was estimated using the linear model y(t) = a + bt, whereby y is the growth rate, t is the ambient temperature, and intercept (a) and slope (b) are the model parameters (Agbessenou et al., 2021). However, in extreme temperatures, the growth rate cannot be accurately captured by a linear function (Vidal et al., 1997). Hence, we selected two empirical nonlinear models (Lactin 1 and CTMI) based on their capacity to explain and predict at least three biologically significant and physically comprehensible namely minimum, optimum, and maximum thermal thresholds as well as maximal growth response of the EPF at specific temperatures (Omuse et al., 2022). The selected nonlinear models were:

Lactin 1 (Lactin et al., 1995):

CTMI (Rosso et al., 1995):

In these models, μm stands for our absolute radial growth (mm day⁻¹) at a particular temperature (T), Tmin denotes the theoretical minimum temperature, Topt denotes the ideal temperature, and Tmax denotes the maximum temperature. In Equations 1, 2, the parameter a is an empirical constant (Lactin et al., 1995).

Data on conidial germination were analyzed using a generalized linear model (GLM), which was predicated on the assumption of a binomial distribution with the log link function. Using Abbott’s formula (Abbott, 1925), the cumulative per cent mortality was adjusted to factor in mortality in the controls, and their normality was tested using the Shapiro–Wilk test (Shapiro and Wilk, 1965) followed by analysis using poisson in a GLM for binomial distribution. Subsequently, the data underwent one-way analyses of variance (ANOVA). The dose.p function was utilized to analyze all time-mortality data with GLM to estimate LT₅₀ estimates and the slopes of the regression curves. The resulting LT₅₀ values and their corresponding slopes were further analyzed via ANOVA. For the analysis of G. pallidipes conidia horizontal transmission and retention data, a GLM was applied, followed by ANOVA. Tukey’s HSD post-hoc test was employed for pairwise comparisons to distinguish means if significant differences were observed, with alpha (α) set at 0.05. Results were considered significant at a 95% confidence level (p < 0.05).

The survival package (Therneau, 2015) was used to perform Kaplan–Meier survival analyses with the Mantel-Cox log-rank chi-squared test to evaluate the effect of SIT-compatible entomopathogenic fungal isolates on the survival rates of G. pallidipes (both conidia donors and receivers). Since the data were normally distributed (Shapiro–Wilk test: p < 0.05), they were subjected to ANOVA followed by Tukey’s HSD post-hoc test. Furthermore, to investigate whether insect sex influences conidia spread between the G. pallidipes conidia donors (males) and conidia receivers (females), we used the unpaired t-test to compare the median lethal time between conidia donors and conidia receivers.

For the temperature-dependent modelling of the SIT-compatible isolates, absolute radial growth was computed using a linear regression model then was fitted into both linear and nonlinear temperature-dependent models. To obtain weighted least-square estimates of the nonlinear model’s parameters, the nls function was employed for datasets with a non-zero residual sum of squares and the nlsLM function from the minpack.lm package for datasets with zero residual sum of squares (Elzhov et al., 2016). Using the rcompanion package, the Akaike’s Information Criterion (AIC) and adjusted R² (adj. R²) goodness-of-fit metrics of the fitted models were compared (Mangiafico, 2020). The adj. R² was calculated from pseudo R² to account for the degrees of freedom in the fitted models as described by Kvålseth (1985). Hence, the best-fitting model was determined to have the least AIC and the highest adj. R². The nonnest2 package’s implementation of Vuong (1989) non-nested likelihood ratio test led to the statistical establishment of equally well-fitting models (Merkle and You, 2020). Unweighted means were used to present the findings of statistical estimations for the parameters. All statistical data analyses were performed using R software (v. 4.1.0; R Core Team, 2020).

The viability of the EPF isolates used in this study ranged between 82.5 ± 1.95 to 99.75 ± 0.22% after 18 h incubation with no significant difference (df = 10; p = 0.9554) among the isolates (Table 1; Supplementary Figure S1). All the EPF isolates screened against the unirradiated (Non-SIT) teneral males of G. pallidipes showed significant differences (χ² = 6.02, df = 10, p < 0.0001) on the mortality of the unirradiated G. pallidipes males 15 days post-infection (Figure 1A), with M. anisopliae strains ICIPE 30, ICIPE 40, and ICIPE 62 causing the highest mortalities (94.44 ± 2.27%, 94.44 ± 2.27%, and 90.28 ± 2.27% respectively) compared to the other fungal isolates; while M. anisopliae ICIPE 7 caused the least mortality of 43.06 ± 3.00% (Figure 1A). In addition, there was significant variation in the mycosis rates (χ² = 4.58, df = 10, p < 0.0001) among the various EPF-challenged unirradiated mycosed cadavers of G. pallidipes, with mycosis rates ranging between 86.11 ± 2.76 to 96.67 ± 2.27% (Supplementary Table S1).

On the other hand, the mortality rates of the EPF-challenged SIT-treated males of G. pallidipes were significantly different (χ² = 5.34, df = 10, p < 0.0001) among the fungal isolates (Figure 1B; Table 1), where M. anisopliae isolates ICIPE 30, ICIPE 07, ICIPE 18, ICIPE 20, ICIPE 40 and ICIPE 41 recorded the highest mortalities (100 ± 0%, 99.96 ± 0.04%, 99.96 ± 0.04%, 99.96 ± 0.04% and 99.96 ± 0.04% respectively); while M. anisopliae ICIPE 32 caused the lowest mortality of 59.94 ± 1.45% 15 days post-infection (Figure 1B). However, there was no significant difference (χ² = 1.75, df = 10, p = 0.5062) in the mycosis rates of the EPF-challenged SIT-treated cadavers of G. pallidipes among the fungal isolates, and the mycosis rates ranged between 91.48 ± 1.42–100 ± 0.0% (Supplementary Table S1).

There was no significant difference (χ² = 2.78, df = 10, p = 0.9861) in the median lethal times (LT₅₀) among the various screened EPF against the unirradiated G. pallidipes with the LT₅₀ values ranging between 8.2 ± 0.9 to 12.5 ± 4.7 days (Figure 1C). However, there was significant difference (χ² = 47.35, df = 10, p < 0.0001) in the median lethal times among the various screened EPF isolates against irradiated G. pallidipes. Metarhizium anisopliae strains ICIPE 20, ICIPE 32, ICIPE 41, ICIPE 62, ICIPE 78 and B. bassiana ICIPE 603 recorded the highest LT₅₀ values (13.3 ± 2.3, 30 ± 6.4, 18.5 ± 5.4, 13.4 ± 4.4, 17.5 ± 10.8 and 11 ± 2.3 days, respectively) and were selected as suitable candidates for compatibility/tolerance with SIT (Figure 1D; Table 1). However, M. anisopliae ICIPE 32 was not used in temperature-dependent assays, modelling, conidia acquisition and transfer bioassays due to the very long median lethal time it had (30 ± 6.4 days) (Table 1). This long median lethal time could be an indication of slower activity against the target pest or maybe the EPF (ICIPE 32) is less adapted to the tsetse’s environment, which may limit its ability to establish infection, or the tsetse (host) may have mechanisms that enable it to evade infection allowing the pathogen to persist within the host for an extended period. Conversely, M. anisopliae ICIPE 40 (despite having a high mortality rate of 99.96 ± 0.04%; Table 1) was not also used in temperature-dependent assays, modelling, conidia acquisition and transfer bioassays due to its relatively short median lethal time (9.3 ± 1.4 days) (Table 1). Moreover, Mutika and Parker (2014) demonstrated that the average longevity of sterilised and starved G. pallidipes males was ≥17 days hence any EPF with LT₅₀ values of more than 20 days and lower than 10 days would not be suitable for SIT.

The interaction of the radial growth of the SIT-compatible EPF, M. anisopliae strains (ICIPE 20, ICIPE 41, ICIPE 62, ICIPE 78) and B. bassiana ICIPE 603 at the different temperature regimes (15–40°C) indicated significant affect the temperature on the radial growth parameters of these strains (df = 20, F = 3.93, p = <0.001) (Figure 2). Among the strains, M. anisopliae ICIPE 78 exhibited the fastest conidia growth at 25°C measuring 2.19 ± 0.16 mm/day, while the slowest growth (0.30 ± 0.0 mm/day) was observed in M. anisopliae ICIPE 62 at 40°C.

The germination rates across the set temperatures for the selected isolates differed significantly (χ² = 165.42, df = 20, p < 0.0001) (Table 2). In addition, the spore yield of the SIT-compatible EPF strains varied significantly across the various temperatures (χ² = 112.82, df = 20, p < 0.0001), with the lowest yield observed at 40°C while the highest yield was obtained at 25°C (Table 3).

The fitted models for the absolute radial growth of the SIT-compatible EPF are depicted in Figures 3A,B. The cardinal temperature model with inflection (CTMI) model (Figure 3A) predicted lower temperature thresholds (Tmin) of 9.0°C, −3.7°C, 8.1°C, 13.1°C and 9.8°C for M. anisopliae ICIPE 20, ICIPE 41, ICIPE 62, ICIPE 78 and B. bassiana ICIPE 603, respectively (Table 4). The upper thresholds (Tmax) were 35.8°C, 35.2°C, 35.6°C, 40.2°C and 45.4°C for M. anisopliae ICIPE 20, ICIPE 41, ICIPE 62, ICIPE 78 and B. bassiana ICIPE 603, respectively (Table 4). Furthermore, the optimum thresholds (Topt) were 29.5°C, 30.8°C, 27.5°C, 26.7°C and 31.1°C for M. anisopliae ICIPE 20, ICIPE 41, ICIPE 62, ICIPE 78 and B. bassiana ICIPE 603, respectively (Table 4). Moreover, a positive relationship between temperature and absolute radial growth rate was observed (R² = 0.87, R² = 0.99, R² = 0.99, R² = 0.77 and R² = 0.84 for M. anisopliae ICIPE 20, ICIPE 41, ICIPE 62, ICIPE 78 and B. bassiana ICIPE 603, respectively) (Table 4).

On the other hand, the Lactin 1 model (Figure 3B) predicted lower temperature threshold (Tmin) of 9.6°C, 4.0°C, 5.3°C, −1.1°C and − 1.8°C for M. anisopliae ICIPE 20, ICIPE 41, ICIPE 62, ICIPE 78 and B. bassiana ICIPE 603, respectively (Table 4). The upper thresholds (Tmax) were 36.0°C, 35.8°C, 35.6°C, 40.0°C and 43.7°C for M. anisopliae ICIPE 20, ICIPE 41, ICIPE 62, ICIPE 78 and B. bassiana ICIPE 603, respectively (Table 4).

Only M. anisopliae ICIPE 20 and ICIPE 78, which emerged as the most thermal stable EPF for incorporation with SIT were used in the horizontal transmission, spore retention and G. pallidipes survival bioassays. Consequently, when the irradiated-EPF-exposed G. pallidipes males were used as conidia donors, there was significant difference in spore acquisition and retention among the conidia receivers (G. pallidipes females) (df = 11, p < 0.001) (Figure 4A). Additionally, 5 days post-exposure to the irradiated EPF-challenged males (conidia donors), the females (conidia receivers) still retained a high quantity of spores (> 1.6 × 10⁶ conidia/mL) for horizontal transfer among their conspecific species. Similarly, when the G. pallidipes females were used as conidia donors and the irradiated G. pallidipes males as conidia receivers, there was significant difference in the conidia uptake and retention among the conidia receivers (irradiated G. pallidipes males) (df = 11, p < 0.001) (Figure 4B). Moreover, 5 days post exposure to the EPF-exposed females, the irradiated males still had sufficient spores (> 0.48 × 10⁶ conidia/mL) to transmit horizontally to their mates. Both the G. pallidipes males (conidia donors) and females (conidia receivers) were susceptible to the selected SIT-compatible EPF isolate infection and there was no significant difference (F = 1.79, Dfn = 3, Dfd = 3, p = 0.6438) attributed to the sex of the G. pallidipes.

When the irradiated-EPF-exposed G. pallidipes males were used as conidia donors and the virgin G. pallidipes females were the conidia receivers, there was no significant difference in the survival rates of conidia receivers (unirradiated G. pallidipes females that had received conidia from EPF exposed irradiated males) (Log-rank (Mantel-Cox) test, χ² = 5.447, df = 2, p = 0.0657) (Figure 5A). Similarly, there was no significant difference in the survival rates of the irradiated-EPF-exposed G. pallidipes males (conidia donors) (Log-rank (Mantel-Cox) test, χ² = 3.86, df = 2, p = 0.1451) (Figure 5B).

On the other hand, when EPF-exposed G. pallidipes females were used as the conidia donors and the irradiated EPF-free G. pallidipes males as conidia receivers, there was no significant difference in the survival rates of the conidia receivers (irradiated G. pallidipes males that had received conidia from EPF exposed virgin females) (Log-rank (Mantel-Cox) test, χ² = 3.82, df = 2, p = 0.1481) (Figure 5C). However, there was significant difference in the survival rates of the EPF-exposed G. pallidipes females (conidia donors) (Log-rank (Mantel-Cox) test, χ² = 7.22, df = 2, p = 0.0271) (Figure 5D). Overall, there was no statistically significant difference in survival rates between the conidia donors and receivers as influenced by the SIT-compatible EPF strains (Log-rank (Mantel-Cox) test, χ² = 11.96, df = 9, p = 0.2153).

There was no sex dimorphism observed in males and females of G. pallidipes after the acquisition of conidia (Log-rank (Mantel-Cox) test, χ² = 11.96, df = 9, p = 0.2153). Furthermore, there were no significant difference in the longevity of the G. pallidipes conidia donors and conidia receivers (both male and female) (t = 0.42, df = 6, p = 0.6888) (Figure 6).