Mating compatibility among Bactrocera cucurbitae (Diptera: Tephritidae) populations from three different origins


Preeaduth Sookar (corresponding author), Entomology Division, Agricultural Services, Ministry of Agro Industry, Food Production & Security, Reduit, Republic of Mauritius. E-mail:


Distinct host ranges of the melon fly, Bactrocera cucurbitae (Coquillett), have been reported among different island populations, suggesting significant genetic divergence. Thus, for the application of the sterile insect technique (SIT), it is important to ensure that released flies are sexually fully compatible with each other and with the laboratory strains. Mating tests among the following strains of the melon fly, B. cucurbitae: Mauritius laboratory-adapted (35 generations), Seychelles laboratory-adapted (24 generations), and Hawaii genetic sexing strain (90 generations), were conducted in field cages at the FAO/IAEA Agriculture and Biotechnology Laboratories in Seibersdorf, Austria during the months of August/September 2009. The genetic sexing strain, developed in Hawaii in 2001, allows separation of females and males on the basis of pupal colour. Two separate series of trials were run simultaneously. In the first, melon fly females from Mauritius were the target strain and the competing males were from Mauritius, Seychelles, and Hawaii (GSS). In the second trial, melon fly females from the Seychelles selected among competing males from the same three populations. Sexual activity was similar among the melon fly populations and no significant non-random, assortative mating was observed. Therefore, it is suggested that melon flies from Mauritius, Seychelles and the Hawaii are compatible, at least under semi-natural conditions.


The melon fly, Bactrocera cucurbitae (Coquillett) (Diptera: Tephritidae), is a serious pest of economic importance attacking fruits and vegetables, especially cucurbits, throughout Asia and the Pacific (White and Elson-Harris 1994), and more recently Africa (Vayssières et al. 2007). Bezzi (1913) first reported it in India, which is considered to be its native home (Dhillon et al. 2005). Bactrocera cucurbitae was discovered in Hawaii (19°34′N, 155°30′W) in 1895 (Back and Pemberton 1917) where it has since been recovered from more than 80 different host plants, including tomato Lycopersicon esculentum Mill., pepper Capsicum frutescens L., watermelon Citrullus lanatus (Thunb.) Mats, cantaloupe Cucumis melo L., pumpkin Cucurbita maxima Duch., cucumber Cucumis sativus L., squash Cucurbita pepo L., bean Phaseolus vulgaris L., eggplant Solanum melongena L., and passion fruit Passiflora edulis Sims (Culliney 2002). However, in the Indian Ocean, both in Mauritius (20°30′S, 57°33′E) and the Seychelles (4°35′S, 55°40′E) it infests only cucurbit crops. It was probably introduced into Mauritius around the end of the 19th century and the beginning of the 20th century (Orian and Moutia 1960) whereas in Seychelles it was observed for the first time in 1999 (Quilici et al. 2001). The main hosts in Mauritius and Seychelles include: cucumber Cucumis sativus L., melon C. melo L., watermelon Citrullus lanatus (Thunb.) Matsun & Nakai, squash C. pepo L., pumpkin C. maxima Duchesne ex Lam., calabash Lagenaria siceraria (Mol.) Stand, ridge gourd Luffa acutangula (L.) Roxb., bitter gourd Momordica charantia L. and snake gourd Trichosanthes cucumerina L. (Quilici et al. 2001).

Control of tephritid pests such as melon fly has traditionally been carried out by chemical means, such as protein bait sprays or male annihilation using cuelure (Steiner et al. 1965). More recently, the sterile insect technique (SIT) has gained wider use against these pests as part of an integrated approach because of its environment-friendly nature (Enkerlin 2005). An example of its effective application is the successful eradication of B. cucurbitae from all of Japan (Kakinohana et al. 1990; Koyama et al. 2004). More recently, Mau et al. (2007) used a genetic sexing strain for melon fly suppression, developed by McInnis et al. (2004), as a component of an area-wide SIT programme in Hawaii.

The success of the SIT in suppressing insect pests depends on sterile male sexual competitiveness in mating with wild females (Knipling 1955, 1979; Dyck et al. 2005). That sterile males are compatible with wild females of the target population is the first pre-requisite. Complete compatibility occurs when there is random mating between two strains. For species distributed in many countries, such as melon fly, and with apparent host range discrepancies between different island populations, it is important to ensure that they are fully compatible and have not developed incipient behavioural differences. Furthermore, unlike the wild populations, which are exposed to the natural environmental conditions, strains reared under laboratory conditions are normally exposed to fairly stable environmental conditions. This can lead to behavioural changes in laboratory strains that may have mating incompatibility with wild strains (Cayol 2000).

The benefits of the release of only sterile males, achieved in the Mediterranean fruit fly, Ceratitis capitata (Wiedemann), as a consequence of the development of genetic sexing strains (Franz et al. 1996; Rendón et al. 2004), has resulted in the extensive use of only a small number of strains worldwide (Dowell et al. 2000; Barnes et al. 2004; Rendón et al. 2004). This has been possible because of high levels of mating compatibility among worldwide populations of this species (Cayol et al. 2002). In the absence of such compatibility, no extensive use of a few strains would be possible. The potential existence of a sibling species complex within B. cucurbitae would impose some difficulties to the implementation of any SIT programme against this pest species. In particular, limiting the use of the genetic sexing strain originated with wild material from Hawaii. The aim of the present work was to evaluate under semi-natural conditions the extent of mating compatibility of B. cucurbitae populations, derived from three different island sites namely Mauritius, Seychelles and Hawaii.

Materials and Methods


The genetic sexing strain of the melon fly developed by USDA ARS, Hawaii (McInnis et al. 2004) was used for the experiment. The colony was held at the FAO/IAEA Agriculture and Biotechnology Laboratories in Seibersdorf, Austria and maintained on a wheat-based diet modified from the standard Seibersdorf diet (Hooper 1987). Flies were kept in Plexiglass cages (50 cm × 50 cm × 60 cm) containing 200 ml pupae and provided with water and enzymatic yeast hydrolysate mixed with sugar in the ratio of 1 : 3. The room was maintained at 25 ± 1°C, 50 ± 10% RH and a photoperiod of 14 : 10 (L : D). Eggs were collected in 1-l polypropylene plastic bottles with evenly spaced oviposition holes of 0.3–0.5 mm. Commercial guava juice (Rubricon, Rubricon Products, Middlessex, UK) smeared inside each bottle served as an oviposition stimulant. The egging device was kept inside the adult cage for 4 h of egg collection. The device was rinsed with water and the volume of eggs measured. The eggs were then transferred to a bottle containing 3 l of water. The bottle with eggs was placed in a water bath (DT Hetotherm, Heto, Denmark) at 28 ± 1°C. The eggs were bubbled for 24 h to provide oxygen. Eggs were then seeded onto kitchen paper (Bounty, SCA Hygiene Products GmbH, Austria) and placed on the standard wheat based diet at the rate of 0.8 ml/kg diet. Fibre glass trays (75 cm × 38 cm × 17 cm) with 4 kg diet were used. After seeding, the trays were brought to the initiation room (25 ± 1°C and 85–90% RH) and covered with a thin cloth and then wrapped with plastic for the first 2 days. On the third day, the plastic was removed. The trays with developing larvae were transferred to the maturation room (22 ± 1°C and 80 ± 5% RH) on day 4. Mature larvae were allowed to jump into sawdust. Pupae were collected for a period of 5 days. They were sieved and kept in the dark in a room at 19 ± 1°C and 60–70% RH for about 10 days. The pupae were then placed in adult cages for emergence.

Melon fly from the Seychelles was brought to the laboratory in Seibersdorf in 2007 in the form of pupae collected from cucurbit fruits from fields in the Seychelles. The colony was maintained for 24 generations under similar conditions as those for the melon fly genetic sexing strain described above, although pumpkin slices were used for collecting eggs for 4 h. The pumpkin slice with eggs was incubated for 24 h in a room kept at 26 ± 1°C in a sealed plastic container lined with wet tissue paper to create a high RH. The pumpkin slice with developing larvae was then placed on an artificial diet similar to that for the genetic sexing strain and the same rearing procedures were followed.

Melon fly from Mauritius was brought to the laboratory in Seibersdorf as pupae in June 2009 from a strain which has been maintained for ca. 35 generations in the laboratory in Reduit, Mauritius since 2006. The rearing procedures were similar to those for the genetic sexing strain with the exception of egg collection. A slice of pumpkin was placed in the egging bottle which was then smeared with fresh pumpkin juice.

Mating compatibility tests

Pupae from the different populations were placed in emergence cages, and every 24 h adults were removed, sorted by sex, and placed in cages with adult food (3 : 1, sugar/hydrolysed yeast) and water until sexual maturation (about 12 days old). One day prior to the experiment, flies were marked to identify their origin. This was performed by placing approximately 10 flies in a mesh bag (1 mm mesh diameter) where, one at a time, they were gently immobilized and painted on the thorax with a dot of water-based paint (DEKA, Unterhaching, Germany) following the standard procedure of the FAO/IAEA/USDA (2003) product quality control manual. Colours, such as green, red and white were interchanged among populations sequentially each day to eliminate the colour effect on the experimental design. After being marked, 20 flies were placed in 1-l containers with the adult food and water and held under laboratory conditions for use the following day.

In each field cage (4 m2 base and 1.8 m in height), one potted Citrus limon (L.) tree, ca. 1.7 m in height with about 1 m diameter canopy, provided the flies with surface for resting and mating activities. Because of unsuitable weather during the experimental period, tests were performed in a greenhouse with controlled temperature (24–29°C) and humidity (60–80%). These laboratory-field cages were identified by number, and each day treatments and observers were randomly assigned to them. The first trial was conducted to assess the compatibility of female melon flies from Mauritius with males from Mauritius, Seychelles and Hawaii. In a second set of trials, the compatibility of female melon flies from Seychelles with males from the three populations was studied. In both trials, a total of eight replicates for each female type were performed. There were 20 males 12, 13, 14 or 15 days old from each of the three populations and 20 females 15 days old of the target strain in each cage for the first four replicates. In the last four replicates, there were 30 males from each of the three populations and 30 females of the target strain. In any one replicate, all the males were of the same age. There were four replicates with males of each of the following ages: 12, 13, 14 or 15 days old. As melon fly mate at dusk, the observation period was from 17.00 to 20.00 hours. Males were released 30 min before females (at 16.30 hours) to allow establishment on the field-caged tree. Mating pairs were collected into 20 ml plastic vials and the time of the mating initiation was recorded. The latency (pre-copulatory period) was defined as the time from the release of the females into the field cage to the beginning of a given copulation.

Statistical analysis

The degree of sexual activity for each strain was determined by the percentage of mating couples obtained in relation to the total number of females released. The mating percentage was transformed to arcsine square-root to normalize the data. The latency data and the number of mating couples in relation to age of flies were subjected to logarithmic transformation before one-way anova. Tukey’s honestly significant difference (HSD) multiple comparison test (post hoc) was then employed to determine significant differences among the means. All analyses were performed using the SAS, Version 8.2 statistical package (SAS Institute, 2001).


The mating percentage and latency to mate are presented in table 1 for Mauritius and Seychelles females. The high participation of females in mating (>93%) showed the suitability of the flies and the environmental conditions for the tests (FAO/IAEA/USDA, 2003). Both types of females showed mating compatibility and random mating with males from the three populations: Mauritius, Seychelles or Hawaii genetic sexing strain (F2,23 = 1.69, P = 0.2092, and F2,23 = 3.26, P = 0.0587, respectively for Mauritius and Seychelles females). There were also no significant differences in the latency among the populations of male melon flies in the presence of females from either Mauritius (F2,168 = 0.52, P = 0.5930) or Seychelles (F2,172 = 0.59, P = 0.5552). However, when data of both trials were compared, females from Mauritius had a significantly lower latency, taking less time (36.9 ± 2.8 min) to copulate than females from Seychelles (44.0 ± 2.5 min) (F1,341 = 16.22, P < 0.0001).

Table 1. Mating percentage (±SE) and latency (±SE) of mating pairs for the melon fly Bactrocera cucurbitae in field cage tests (eight replicates) with females from either Mauritius or Seychelles when released together with males from Mauritius, Seychelles and Hawaii genetic sexing strain
♂Origin Mauritius ♀Seychelles ♀
Matings (%)Latency (min)Matings (%)Latency (min)
  1. Within column, mean values followed by the same letters are not significantly different (Student–Newman–Keuls, P = 0.05). Data on mating % and latency were arcsine transformed and log transformed, respectively, before analysis.

Mauritius33.4 ± 1.7 a47.7 ± 5.9 a32.0 ± 2.9 a49.0 ± 5.0 a
Seychelles26.4 ± 3.5 a31.7 ± 4.2 a37.6 ± 4.8 a43.6 ± 3.5 a
Hawaii GSS32.9 ± 3.6 a30.4 ± 4.0 a23.6 ± 3.8 a37.7 ± 4.1 a

The mean numbers of pairs for each male mating age (table 2) were not significantly different (F3,11 = 2.85, P = 0.1052 and F3,11 = 0.08, P = 0.9691), respectively for matings with Mauritius and Seychelles females. Older male flies (14 and 15 days old) took significantly less time (F1,168 = 50.91, P < 0.0001) to start to copulate (latency) when compared with younger male flies (12 and 13 days old) in the presence of females from Mauritius. In the case of Seychelles females, 15 days old males took significantly less time (F1,172 = 12.77, P < 0.0001) to copulate when compared with 12 days old males.

Table 2. Mean number (±SE) and latency (±SE) of mating pairs for the melon fly Bactrocera cucurbitae in field cage tests (eight replicates) with females from either Mauritius or Seychelles when released together with 12, 13, 14 or 15 days old males from Mauritius, Seychelles and Hawaii genetic sexing strain
Age of ♂Mauritius ♀Seychelles ♀
No. pairsLatency (min)No. pairsLatency (min)
  1. Within column, mean values followed by the same letters are not significantly different (Student–Newman–Keuls, P = 0.05). Data on mean number of pairs and latency were log transformed before analysis.

1212.3 ± 1.4 a69.7 ± 7.1 a12.7 ± 0.7 a73.3 ± 5.8 a
1312.0 ± 1.7 a64.2 ± 4.6 a13.3 ± 3.0 a39.1 ± 4.1 b
1415.7 ± 0.7 a9.6 ± 1.1 b16.0 ± 4.7 a23.7 ± 2.4 c
1516.3 ± 0.9 a18.1 ± 2.4 b15.7 ± 5.2 a43.7 ± 4.2 b


The present study analysed the sexual compatibility of melon fly populations from three different origins around the world, namely Mauritius, Seychelles and Hawaii. The data demonstrate that from a qualitative standpoint, there has not been any change in mating performance and latency of the flies from the three populations despite years of geographical isolation. It was also found that the ‘wildish’ strain from Seychelles was compatible with the genetic sexing strain from Hawaii. Even though there was no effect of male age in terms of mating pairs formed, older male flies (15 days old) took less time to copulate when compared to younger males (12 days old) in the presence of females from either Mauritius or Seychelles.

Long periods under mass-rearing conditions can adversely affect the performance of sterile fruit flies (McInnis et al. 1996; Cayol 2000). Wong et al. (1982) reported that laboratory reared oriental fruit fly, Bactrocera dorsalis Hendel, cultured for ca. 330 generations preferred to mate in field cage experiments with members of its own strain, whereas a wild strain also preferred to mate with members of its own strain. In this study, there was no sexual isolation observed among Mauritius or Seychelles females and any of the three male strains despite various periods under mass rearing.

The genetic sexing strain melon fly was found to be effective in suppressing the wild melon fly population in Hawaii (Mau et al. 2007). This was achieved with a relatively low sterile wild fly ratio compared with other strains with other species of fruit flies, indicating that the sterile strain was very competitive in the field (McInnis et al. 2007). As the melon flies from Mauritius and Seychelles are sexually compatible with the genetic sexing strain, the latter could be used in a SIT programme to suppress the population of melon flies in either Mauritius or Seychelles. This information is of great relevance for authorities in either Mauritius or Seychelles to determine the feasibility of using the genetic sexing strain from Hawaii as part of an integrated area-wide suppression programme.

This study indicates that there is no evidence for incipient speciation, much less for the existence of a sibling species complex, within B. cucurbitae, even though there are distinct differences reported in the host range between the populations assessed. In C. capitata, there are also differences in host ranges among different populations, nevertheless full mating compatibility has been found among them (Cayol et al. 2002). Of course, to confirm the results of the present B. cucurbitae study, it needs to be expanded to include other populations from the geographic extremes of the Asia/Pacific region and also the African mainland.


This study was funded by the FAO/IAEA through the research contract No 12859. We would like to thank Mr A. Islam and Mr T. Dammalage for their support in the laboratory, as well as Teresa Vera (Estación Experimental Agroindustrial Obispo Colombres, Tucumán, Argentina), Jorge Hendrichs and Rui Pereira (both from, Joint FAO/IAEA Division of Nuclear Techniques in Food and Agriculture, Vienna, Austria) for critical reviews of an earlier version of this manuscript.