The study was conducted over a period of 10 months (from August 2019 to May 2020) in the laboratory. During this period, the average daily temperature varied from 23.05˚C (May) to 27.42˚C (December), and the relative humidity varied from 45.24% (February) to 76.48% (August). The photoperiod was maintained at 12 hours of light a day.

The first diet (local Diet 1) consisted of several protein sources namely soybean meal, rice bran, corn flour and yeast waste. According to [10] , soybean meal and rice bran contain respectively 44.46% and 13.25% of protein. Corn flour contains 6.05% of protein [11] . Yeast waste was boiled for 30 minutes before mixing with other components. The diet obtained had a pasty appearance. The second diet (Diet 2) was obtained by mixing different components in water. This diet was liquid and contains two types of yeast: LBI2240 and FNILS65. LBI2240 is a debittered brewer’s yeast without the use of chemicals and is deactivated by being heated and roller dried. LBI2240 is a whole-cell yeast while FNILS65 is hydrolysed yeast. LBI2240 and FNILS65 contain respectively 45.40% and 65% of protein [12] . Table 1 presents the composition of the two diets used to rear B. dorsalis larvae.

The B. dorsalis larvae were reared in the laboratory under the same temperature and relative humidity conditions of the experiment room. Eggs were first collected by placing artificial nests in rearing cages containing a population of sexually mature B. dorsalis. This population was derived from a strain of B. dorsalis bred since 2018. Each artificial nest consisted of a yellow funnel perforated with small equidistant 1-mm-diameter holes and containing black muslin cloth moistened with mango juice. After 24 hours of exposure, the nests were removed and rinsed with tap water. This water was then filtered through a very-fine-mesh muslin cloth to collect the eggs.

The collected eggs were then placed using a brush in plastic pots containing the diets. For Diet 1, eggs incubation was carried out on blotting paper previously placed on the diet. For Diet 2, the eggs were spread on vegetable sponges previously placed in the substrate. Observations were made daily to monitor the hatching of the eggs and the progression of the different larval stages (L1, L2, L3) that feed on diets. First-instar larvae (L1) are recognizable by their egg-like whitish colour and poorly developed mouth stylets. Second-instar larvae (L2) are

intermediate larvae. They begin to take on the colour of the nutrient medium and are slightly larger in size than the first-stage larvae. The segments become increasingly visible. Third-instar larvae (L3) are well developed, are able to jump, and the segments are clearly visible. At each observation, Diet 1 was soaked in distilled water to prevent it from drying out. These third-stage larvae were recovered 6 to 7 days after incubation and then placed in jars containing sand sterilized at 100˚C for 12 hours. These jars were covered with a muslin canvas held by elastic bands.

The pupae formed at the end of the larval cycle were finally collected by sieving the sand. They were placed in Petri dishes and transferred to rearing cages for the emergence of adult flies.

Microsoft Office Excel was used for data entry and processing. Data processing consisted of applying the above formulas before performing the statistical analyses. Statistical analysis was conducted using R software 3.6.0. The Kruskal-Wallis and Wilcox comparison tests were used to compare the means in case of significant difference at the probability threshold of 5%.

Table 2 presents the effect of diet on the duration of each developmental stage of B. dorsalis. Egg hatching lasted an average of 1.95 days on Diet 1 and 2.15 days on Diet 2. No significant difference (W = 0.84; P = 0.61) was observed between these two diets for this stage. Similar results were recorded for the duration of development of the different larval stages. First-instar larvae reached the second instar after 2.94 days on Diet 1 and 3.30 days on Diet 2. Second-instar larvae reached the third instar after 1.85 days on Diet 1 and 2.03 days on Diet 2. Third-instar larvae pupated after 1.97 days and 2.23 days on Diet 1 and Diet 2, respectively. The duration of pupal development was significantly influenced (W = 0.97; P = 0.003) by the diet. The pupae started to hatch after 8.51 days for Diet 1 and 10.28 days for Diet 2.

Table 3 presents the survival rate of the different developmental stages. After incubation, 32.12% of the eggs hatched on Diet 1, while 74.87% of the eggs hatched on Diet 2. The survival rate of the larvae in the different developmental stages varied significantly as a function of diet. From the first-instar larvae obtained after egg hatching, 98.75% and 85.52% were able to reach the second instar on Diet 1 and Diet 2, respectively. The survival rate of second-instar larvae was 97.75% on Diet 1 and 82.61% on Diet 2. No significant difference (P = 0.7) was observed between the two diets with regard to the survival rate of third-instar larvae. In total, 98.24% and 97.82% of third-instar larvae on Diet 1 and Diet 2, respectively, were able to become pupae. Adult flies emerged from these pupae with an emergence rate of 88.33% on Diet 1 and 95.36% on Diet 2. Statistical analysis revealed a highly significant difference (P = 0.001) between the two food substrates for this developmental stage.

In the same column, the means followed by the same letter are not statistically different at the 5% probability threshold. NS: Not Significant; **: Highly Significant.

In the same column, the means followed by the same letter are not significantly different at the 5% probability threshold. NS: Not Significant; **: Highly Significant; ***: Very Highly Significant.

The emergence of adult flies was spread over 3.12 days for Diet 1 and over 3.56 days for Diet 2 (Table 4). For these two respective diets, average pupal hatching rates of 89.12% and 97.93% were recorded. The flight ability rates were 98.59% and 97.63% for Diet 1 and Diet 2, respectively. No significant difference (W = 83.5; P = 0.095) was observed between the two diets for this parameter. The sex ratio was 1 male to 0.92 females for Diet 1 versus 1 male to 1.10 females for Diet 2. Emergence duration, emergence rate and sex ratio were significantly influenced (P < 0.05) by the type of diet.

The female B. dorsalis experienced a preoviposition period of 3.25 days and 5.81 days after emergence for Diet 2 and Diet 1, respectively (Table 5). The average laying time was 57.75 days for females from Diet 1 and 109.81 days for those from Diet 2. The average daily spawning indices were 4.76 eggs/female/day for Diet 2 and 20.54 eggs/female/day for Diet 1. All these parameters were significantly influenced (P < 0.05) by the type of diet.

Figure 2 illustrates the progression in the number of eggs laid by females of B. dorsalis from Diet 1. The first eggs laid were observed 72 hours after the emergence of the females. The maximum number of eggs laid was recorded between the 8^th and 28^th days. The peak of laying was observed on the 15^th day, with an average of 40 eggs laid per female on that day. No eggs were laid from the 78^th day onwards after emergence of the females.

In the same column, the means followed by the same letter are not significantly different at the 5% probability threshold. NS: Not Significant; *: Significant; ***: Very Highly Significant; M: Male; F: Female.

In the same column, the means followed by the same letter are not significantly different at the 5% probability threshold. ***: Very Highly Significant.

Figure 3 shows the progression in the number of eggs laid by female B. dorsalis from Diet 2. The females started to lay 24 hours after their emergence. The maximum number of eggs laid was observed between the 15^th and 40^th days. The peak of laying was recorded on the 28^th day after emergence, with an average of 14 eggs laid per female on that day. No eggs were laid after the 157^th day.

The lifespan varied on average from 42.02 days to 107.18 days depending on the sex and diet (Table 6). Flies from larvae fed Diet 2 lived longer than those from larvae fed Diet 1. In addition, males lived longer than females regardless of the type of diet. The multiple comparison test showed a very highly significant difference ( X 3 2 = 91.74; P < 0.0001) between the two diets.

In the same column, the means followed by the same letter are not significantly different at the 5% probability threshold according to Kruskal-Wallis multiple comparison test. ***: Very Highly Significant.

Figure 4 shows the progression of the survival rate of adult males and adult females of B. dorsalis according to the two larval diets. In males that emerged from larvae fed Diet 1, approximately 50% of the flies died 60 days after emergence. This survival rate was observed 45 days after the emergence of female flies obtained from Diet 1. For the males and females that emerged from the larvae fed Diet 2, survival rates of 50% were observed 83 days and 63 days after emergence, respectively. The last female and male mortalities were observed 140 days and 145 days after their emergence for Diet 2, respectively. For Diet 1, the last female and male mortalities were observed 81 days and 100 days after their emergence, respectively.

The two food substrates were all favourable to the development of the different stages of B. dorsalis. The type of food substrate did not significantly influence the egg incubation time or the development time of the L1, L2 and L3 larval stages, which varied from 1.85 to 3.30 days. Previous studies have highlighted the life cycle of B. dorsalis. [14] reported a duration of 1.61 to 4.5 days for the development time of B. dorsalis eggs and larvae on three varieties of mangoes. [15] reported an egg incubation time ranging from 2.68 to 3.12 days on fruits of four host plants, including mango. [13] reported egg incubation times of 2.60 and 3.33 days on mango and orange, respectively. On these two respective host fruits, the total duration of development of the L1, L2 and L3 larval stages was 7.97 and 8.73 days, respectively. Conversely, the egg incubation time noted on our two food substrates is similar to that reported by [16] for the same species reared with mango pulp. Pupation duration was significantly influenced by the type of food substrate. It was 8.51 days for food substrate 1 and 10.28 days for food substrate 2. These results are similar to those reported by [14] for mango and [13] for orange. According to [15] , the duration of pupal development varies from 7 to 7.80 days on the following host fruits: guava, banana, mango and sapota. For this parameter, the difference observed between food substrates 1 and 2

is likely be related to the chemical composition of the substrates or to the content of a chemical element that plays a major role in the development of B. dorsalis pupae. [17] showed that a yeast concentration of 20 g/L water results in a shorter development time from egg to pupa in B. dorsalis (11.77 days) than concentrations of 10 g/L, 60 g/L and 100 g/L. In our study, this development time was 8.71 days on food substrate 1 and 9.71 days on food substrate 2. The developmental time of immature stages of B. invadens is also affected by temperature with the duration of each stage decreasing as temperature increased [18] . The daily temperatures (23˚C - 27.5˚C) recorded during our study are included the optimal range of 20˚C - 30˚C reported by these authors.

The survival rate of the different developmental stages of B. dorsalis was significantly influenced by the food substrate. The egg-hatching rate was lower on food substrate 1 (32.12%) than on food substrate 2 (74.87%). When food substrate 2 was liquid, it probably maintained a relative moisture level that favoured egg hatching. Food substrate 1 was in a pasty state, so distilled water was sprinkled on it during the observations to prevent it from drying out. Conversely, the survival rate of larvae and pupae (82 to 99%) was higher on food substrate 1 than on food substrate 2. Under the same rearing conditions, [19] reported larval survival rates of 49% to 85.4% depending on the mass of food substrate 1 (30 g/100 larvae, 50 g/100 larvae, 70 g/100 larvae, 90 g/100 larvae, 110 g/100 larvae) used to feed B. dorsalis larvae. [17] highlighted the role of yeast in the development of B. dorsalis. These authors reported that B. dorsalis cannot survive until the adult stage when the artificial food substrate fed to the larvae does not contain yeast. As part of their study, yeast concentrations of 10 g/L water and 20 g/L water in the food substrate of the larvae made it possible to obtain pupation rates of 82% and 88%, respectively. With these two respective concentrations, 78.71% and 78.17% of the pupae hatched. Beyond these concentrations, the pupation rate and the adult emergence rate decreased significantly. The pupation rates recorded on food substrate 1 and food substrate 2 were 98.24% and 97.82%, respectively. This performance would be linked to a good concentration of yeast used in the two food substrates. These pupation rates are well above those reported on mango (74.17%) and orange (35%) by [13] .

The emergence rate of B. dorsalis adults was 89.12% for food substrate 1 and 97.93% for food substrate 2. This significant difference between the two food substrates lies in the content of certain chemical elements, such as proteins, which mainly make up the two types of yeast used. [17] reported a decrease in the emergence rate of B. dorsalis when the amount of yeast increases in the food substrate fed to larvae. [20] indicated that the emergence rate of adult flies is not influenced by the type of yeast or the structure of the food substrate. In this study, the sex ratio was 1 male to 0.92 female for substrate 1 and almost identical for substrate 2. According to [13] , the sex ratio of B. dorsalis is 0.74 on mango and 0.87 on orange. The flight ability of B. dorsalis was not influenced by the larval food substrates. The result was similar to those obtained by [12] who observed 98.80% of adult fliers in B. dorsalis by rearing the larvae with a mixture of LBI2240, FNILS65 and Wheat germ oil.