Introduction

Bactrocera cucurbitae (Coquillett) is a tephritid fruit fly native to the Indo-Malayan region.¹ Its distribution has extended to include Africa, temperate Asia, and a number of Pacific Islands.² It became established in Japan in 1919, where it was first detected in the Yaeyama Islands, the most southern islands. Melon fly subsequently spread north in the Southwestern Islands of Japan to the Miyako Islands (1929), Kume Island (1970), the Okinawa Islands (1972), Yoron and Okierabu Islands (1973), Tokunoshima Island, the Amami-Oshima Islands, and Kikai Island (1974). The melon fly also subsequently spread to the Daitoh Islands (1977) (Fig. 1).³ Melon fly is a very serious pest of cucurbit crops, but also attacks fruits in a number of other plant families.² Adult female melon flies lay eggs in these fruits, potentially even before the flower opens. Subsequent larval feeding can cause considerable fruit damage that ruins the crop for local consumption and necessitates the development of quarantine protocols to prevent the introduction of this pest to other areas where this pest is not established.¹

Figure 1.

Japan, with enlargement of the Southwestern Islands of Japan, where the successful melon fly eradication campaign was conducted. Outline maps of Japan and Taiwan modified from world map clipart source,⁴⁸ with enlarged Southwestern Islands of Japan obtained from GADM⁴⁹ and incorporated into the final figure using ArcGIS.⁵⁰

In 1972, the Ministry of Agriculture, Forestry and Fisheries (MAFF) of Japan and the Okinawa Prefectural Government (OPG) initiated an experimental melon fly eradication program based on the use of sterile insect technique (SIT) in one of these islands (Kume Island). Following successful eradication there in 1978, eradication efforts were extended to the Miyako, Okinawa, and Yaeyama groups of Islands, in 1984, 1986, and 1989, respectively, with eradication achieved in 1987, 1990 and 1993, respectively.⁴ Overall, this eradication program is one of the highlights in the history of melon fly control. Over the course of this successful program, many fruits were collected and held for the assessment of infestation by melon fly. The infestation data were published in a number of in-country reports, all written in Japanese, which have not been readily accessible by people outside of Japan. As part of an effort to develop a worldwide database on the status of fruits as hosts of melon fly, we reviewed and summarized the fruit infestation data gathered from host fruits collected before the initiation of suppression activities (bait sprays, male annihilation, and sterile fly releases). This summarization is of particular value because of the quantity of data collected and the fact that there is no longer any infestation of these fruits in Japan; so this paper provides historical documentation of melon fly hosts in Japan. This summarization will also be of value should melon fly ever become re-established on any of the Southwestern Islands of Japan, for use in pest risk assessments and in contributing to our overall understanding of the host range of the melon fly.

Materials and Methods

Data Collection

Throughout the course of the eradication program, results of efforts on the different islands were summarized in government documents published on an annual basis. Agencies responsible for these publications included the South Plant Protection Office, The Miyako Plant Protection Office, the Department of Agriculture, Forestry and Fishery of Okinawa and Kagoshima Prefectures. These publications included data tables that summarized what cultivated and wild species of fruits were collected, the numbers of fruits collected and assessed for infestation by melon fly, and the numbers of fruits found to be infested. These summaries included fruit collections made before the start of suppression efforts, and continued until eradication was declared. Here, we used data only from the fruit collections made before the start of any island-wide melon fly suppression efforts (though individual farmers were utilizing some conventional controls on their farms) in order to present natural infestation rates. The dates of pre-suppression fruit sampling, suppression by bait sprays and male annihilation, and suppression by SIT are given in Table 1 for the islands involved in the eradication program. Methods description of fruit collections made in the course of the eradication program indicates that collected fruits were examined and those that were found to be rotten (which could be a result of tephritid fruit fly infestation) were dissected within 3 days after they were collected. If larvae were detected, they were allowed to continue to develop inside the fruits, with the fruits placed on sawdust or sand in small plastic cups. After 2 weeks, the sawdust or sand was sieved to collect the pupae. Pupae were placed in a Petri dish inside cages until adult emergence.⁵ Fruits that were not found to be rotten were similarly held on sand in plastic containers for 2–3 weeks at ambient temperature, after which time containers were checked for pupae and/or adult insects, with pupae held as described above until adult emergence.⁶ Species identification of emerged adults was based on reference to Drew (1989).⁷

Table 1.

Dates of fruit collections, by Island or Island group, taken before the start of melon fly population suppression, as well as dates of bait spray and male annihilation (MA), dates of sterile insect technique (SIT), and dates of declaration of melon fly eradication.

Data Presentation

Plant species from which fruits were collected are presented in alphabetical order by genus, within an alphabetical listing of plant families. For each plant species listed, common names in both English and Japanese are given, where possible (Table 2). Japanese common names used were based on the Flora of the Ryukyus, South of Amami Island,⁸ with minor exceptions. Scientific names used are, where possible, based on GRIN (Germplasm Resources Information Network) taxonomy for plants.⁹ For species not included in the GRIN database, names used were based on the Flora of the Ryukyus, South of Amami Island.⁸ In cases where current scientific names differ from older names used in publications, the host data are presented under the currently accepted scientific name, with footnotes indicating the scientific name used in the original publications. There were several cases where more than one common name was used for a plant species (see Table 2). In some cases (eg, common names used for Momordica charantia var. pavel Crantz), this just represented different common names used on different Islands, so data from the two names were combined. However, in another case (ie, Capsicum annuum cv. Glossum), although the two common names are considered in the Flora of the Ryukyus, South of Amami Island⁸ to refer to the same species, they are treated in the Japanese market as different commodities, so the data were kept separate. The approach taken (ie, combining data or keeping data separate) for each species for which two common names were used is indicated in Table 2. Data are summarized by five island groups: Amami Islands, Kume Island, Miyako Islands, Okinawa Islands, and Yaeyama islands. Data sources, by Island, are as follows: Amami Islands:^{1011121314151617–18}. Kume Island:¹⁹. Miyako Islands:^{202122232425262728–29}. Okinawa Islands:^{26,27,293031323334353637–38}. Yaeyama Islands:^{26,27,29,3738394041–42}. The total number of fruit collections made, total numbers of collections made where infestation was detected, total numbers of fruits collected, total numbers of infested fruits, and average percentage B. cucurbitae infestation are presented by Island groups for each plant species. The overall average percentage infestation across all island groups was calculated for each collected plant species as an average weighted by the number of fruit collections made in each island group (ie, averages from Islands where more collections were made had greater weight in the overall weighted average; Footnote 2 in Table 3 provides a sample of the weighted average calculation). A collection is defined here as an entry into a summary table. These entries may represent the sum of collections made at a number of different sites, but are typically collections made over the course of 1 month. Although the summary publications sometimes indicate the number of sites from which fruits were collected, this was not always the case, which is why the line/month total was used as an indication of a collection. One deviation from this were the 1979 data for Okinawa, where the published monthly data only included monthly collections where infestation was recovered. A summary table, however, indicated the total number of infested fruits and total number of fruits collected over the course of 1979.³³ The latter data were used for our summarization, because it more accurately represented the infestation rate.

Table 2.

Scientific and common names of fruits collected in the Southwestern Islands of Japan for assessment of infestation by melon fly. Those fruits from which melon fly was recovered are listed first followed by fruits collected from which no melon fly was recovered.

Results

Over 1.1 million fruits were collected before the start of population suppression efforts over the course of the B. cucurbitae eradication program (the sum of fruit numbers listed in Tables 3 and 4). These fruits encompassed 39 different plant taxa, of eight plant families, of which infestation was found in 24 taxa (spread across four plant families) (see Tables 3 and 4). Data on collections of taxa from which melon fly was recovered are summarized in Table 3, with references used for each island or island group noted in a footnote to the table. Included there are the total number of collections made for each plant species, the number of collections where melon fly infestation was found, the total number of fruits collected, the total number of infested fruits recovered, and the overall infestation rate. Data on collections of taxa from which melon fly was not recovered are summarized in Table 4. Infestation was found in four plant families: Caricaceae, Cucurbitaceae, Moraceae and Solanaceae. Most of the host fruit species were in the family Cucurbitaceae (15 species). Solanaceae was the plant family with the second most number of host species (6 species). In addition to fewer recorded solanaceous host species than cucurbitaceous host species, infestation rates were considerably lower in solanaceous hosts than in cucurbitaceous hosts. The identified hosts in the plant family, Moraceae, were both very poor hosts, with infested fruits found in only one fruit collection for each species, despite collection numbers exceeding 12,000 (Ficus pumila) and 34,000 (F. erecta). The highest overall infestation rates came from Cucurbita pepo (33.3%; zucchini squash), but this was based on a rather small sample size (13 fruits). The next highest infestation rate was found in Momordica charantia (21.6%; bittermelon) followed by Trichosanthes ovigera (17.7%; snake gourd), Cucumis melo (17.2%; cantaloupe), Lagenaria siceraria cv. gourda (16.7%; bottle gourd), and Luffa aegyptiaca (15.3%; smooth luffa), all cucurbitaceous crops.

Discussion

Comparing the results presented here with B. cucurbitae host listings reported by other authors,^{43444546–47} there are four new B. cucurbitae host species reported in the data reported herein, which had not previously been reported, as well as several new varieties of host species that had previously not been listed by other authors. The four new host species are Ficus erecta, F. pumila, both in the plant family Moraceae, Solanum erianthum (Solanaceae), and Zehneria liukiuensis (Cucurbitaceae). Earlier host listings for B. cucurbitae had included other Moraceae species, such as Ficus carica L.^45,46 and F. chartacea,^43,44 but we are unaware that the two Ficus species reported here as melon fly hosts have previously been listed as B. cucurbitae hosts outside of the Okinawa publications related to the melon fly eradication program. Infestation rates for these two Ficus species were not high. For F. erecta, only three fruits were found to be infested out of 34,749 collected fruits, and the average percentage infestation rate was only 0.0025%. For F. pumila, only seven fruits were found to be infested out of 12,101 collected fruits, and the average percentage infestation rate was only 0.015%. The overall average infestation rate of S. erianthum was higher than for the Ficus spp. (3.3%), but varied from 0.0% to 8.0% among island groups. The last of the four new host species identified here for melon fly, Zehneria liukiuensis, supported infestation by melon fly in 30.0 to 50.0% of collections, but the average percentage infestation rate was low, only averaging 1.4%. In the 1983 publication reporting on the infestation rate of host fruits of melon fly in the Yaeyama Islands in 1982,²⁷ it was reported that ten Trichosanthes kirilowii Maxim.(Cucurbitaceae) fruits were collected of which one (10%) was found to be infested by B. cucurbitae. This would be an additional new B. cucurbitae host. However, we think that these fruits were incorrectly designated as T. kirilowii fruits because T. kirilowii is listed by the Flora of the Ryukyus to only occur in the Amami Islands and not in the Yaeyama Island group. Additionally, such an error could stem from the close similarity of spelling and sound of the representative common names in Japan: “Ki-karasuuri” (T. kirilowii) versus “Ke-karasuuri” (T. ovigera). Because of the identity question, we have not included these collection results in the summary data presented in Table 3.

Of the plant species listed from which no B. cucurbitae infestation was found (Table 4), infestation has previously been reported by other authors for most of the species. No record of B. cucurbitae infestation has been reported elsewhere for only three of the listed species: Ficus thonningii Blume (Moraceae), Solanum capsicoides All., and S. mammosum L. For most of the fruits collected where no infestation was found, fruit collection numbers were not high (under 1,000 total), with the exception of Phaseolus vulgaris L. (9,537) and Capsicum frutescens L. (17,247). It may be that low-level infestation could have been found in these plant species also if collection numbers had been considerably increased, because all three species are fairly closely related to plant species in which infestation has been reported.

Considering that melon fly infestation can lead to significant crop loss and can necessitate the development of postharvest quarantine treatments to permit exportation of melon fly susceptible crops out of a place of production, the eradication of melon fly throughout the Southwestern Islands of Japan gave a major positive impact to the production of the fly-susceptible crops. This impact, though, was achieved as a result of considerable financial and labor investment. The eradication program was achieved through the application of SIT. An initial step for this was to reduce wild fly populations in order that the released sterile fly numbers significantly exceeded wild fly population numbers. Wild fly population suppression was achieved through the use of male annihilation (traps incorporating a male lure and a toxicant) and protein bait sprays. Facilities for mass rearing of melon flies for sterile fly release were constructed with fly production ramped up to as high as 200 million flies per week. Produced flies then had to be irradiated and dispersed throughout the Southwestern Islands. Ongoing evaluation of the effectiveness of the sterile fly releases also had to be implemented through trapping assessment of sterile versus wild fly numbers, assessing the level of infestation of melon fly-susceptible fruits, and assessing the hatchability of eggs produced by wild female melon flies.³ Clearly, considerable effort was exerted; but considerable benefit was achieved. Now, as long as there is no reinvasion by melon fly, there is one less pest problem to which growers must attend.

Author Contributions

Conceived and designed the experiments: GM, TT. Analyzed the data: GM. Wrote the first draft of the manuscript: GM. Contributed to the writing of the manuscript: GM, TT. Agree with manuscript results and conclusions: GM, TT. Jointly developed the structure and arguments for the paper: GM, TT. Made critical revisions and approved final version: GM, TT. All authors reviewed and approved of the final manuscript.