Biological Control 34 (2005) 144–151 www.elsevier.com/locate/ybcon
Establishment of Bathyplectes anurus (Hymenoptera: Ichneumonidae), a larval parasitoid of the alfalfa weevil, Hypera postica (Coleoptera: Curculionidae) in Japan Megumi Shoubu a,¤, Masami Okumura a, Akinori Shiraishi a, Hidenori Kimura a, Masami Takagi b, Takatoshi Ueno b a
Moji Plant Protection Station, Ministry of Agriculture and Fisheries, Moji Ward, Kitakyushu City, Fukuoka 801-0841, Japan b Institute of Biological Control, Faculty of Agriculture, Kyushu University, Fukuoka 812-8581, Japan Received 15 November 2004; accepted 21 April 2005 Available online 1 June 2005
Abstract The alfalfa weevil invaded Japan in the early 1980s. In Southwestern Japan, the weevil infests Chinese milk vetch, which is a main source of honey products. Since apiarists avoid application of insecticides, four species of parasitoid wasps were introduced from the US into Japan for biological control of the weevil in 1988 and 1989. In 1996, one of the parasitoids, Bathyplectes anurus (Thomson) was recovered. Accordingly, we started the survey to assess the incidence and eVectiveness of this parasitoid in suppressing the alfalfa weevil. B. anurus expanded its distribution during 1998–2003. In 1998 and 1999, the percentages of parasitism were mostly less than 5% but quickly increased to about 40% in 2003. The survey also showed that the extent of damage of the weevil on Chinese milk vetch decreased from 2001 to 2004; there was a negative correlation between the extent of weevil damage and the percentage parasitism one year previously. These results suggest that the parasitoid reduced damage by the alfalfa weevil. 2005 Elsevier Inc. All rignts reserved. Keywords: Bathyplectes anurus; Astragalus sinicus; Chinese milk vetch; Classical biological control; Parasitism; Introduction
1. Introduction The alfalfa weevil Hypera postica (Gyllenhal) is native to the Old World and has a wide distribution there, occurring throughout Europe and ranging to North Africa, the Middle East, India, and Western Asia (Clausen, 1977). The weevil mainly infests legumes (Essig and Michelbacher, 1933). It is rarely of economic importance in its original range (Clausen, 1977; Essig and Michelbacher, 1933). In the US, the weevil was Wrst discovered in Utah in 1904 (Titus, 1910). Then it was found to occur in Ari-
*
Corresponding author. E-mail address:
[email protected] (M. Shoubu).
1049-9644/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.biocontrol.2005.04.009
zona in 1939 (Wehrle, 1940) and Maryland in 1951 (Poos and Bissell, 1953) independently, and it subsequently spread throughout the US. Like other pests that invaded a new area without their natural enemies, the alfalfa weevil became a destructive pest that damaged the production of alfalfa in North America; it was the most important pest of alfalfa in the US (USDA, 1968, 1972). The alfalfa weevil was the target of biological control after its introduction to the US; 12 parasitoids were introduced as candidates for biological control in the states (Brunson and Coles, 1968; Bryan et al., 1993; Chamberlin, 1924a,, 1926). To date, biological control of alfalfa weevil has met great success in Northeastern and North Central States (Day, 1981; RadcliVe and Flanders, 1998).
M. Shoubu et al. / Biological Control 34 (2005) 144–151
In 1980s, the alfalfa weevil was accidentally introduced into Japan. Discovered in Okinawa Pref. and Fukuoka Pref. in 1982, it infested leguminous forbs, including Vicia angustifolia L., Trifolium repens L., Medicago polymorpha L., and Melilotus oYcinalis (L.) (Baba, 1983; Tao, 1984). In Japan, those legumes are “forbs”; they have no commercial value. The weevil was not found to attack farm products in early 1980s. The most suitable host of the weevil, alfalfa, is commercially grown only in Hokkaido, a Northern Island where weevil damage has not been reported. Thus, there was no economic damage for some years after its invasion into Japan (Baba, 1983; Tao, 1984). However, the weevil became a serious pest in Chinese milk vetch, Astragalus sinicus L. after the late 1980s (Morimoto, 1987). Chinese milk vetch had not recorded as a host plant of the weevil until then. It is originally from China (Yasue, 1987), where the alfalfa weevil does not present (Essig and Michelbacher, 1933). Hence, it is likely that Chinese milk vetch is a new host plant for the weevil. Chinese milk vetch is planted in fallow paddy Welds from winter to spring as a green manure crop, and the Xowers are the main source of honey production in Japan. Furthermore, honey collected from Chinese milk vetch is thought to be of the best quality and the most valuable among all kinds of honey (Yasue, 1987). Chinese milk vetch Xowers during the period of weevil larval occurrence in Japan. The larvae of the weevil feed on Xower buds and Xowers (Ono et al., 1987). When large numbers of weevil larvae are present, Xowers of Chinese milk vetch are entirely devoured, and the production of honey from Chinese milk vetch is greatly decreased. Application of insecticides is avoided by apiarists because they may kill bees and possibly contaminate honey products. Accordingly, Moji Plant Protection Station initiated a project of biological control of the weevil. Although there are several indigenous natural enemies in Japan, they are ineVective in reducing populations of the alfalfa weevil (Okumura, 1991; Okumura et al., 1987; Yamaguchi et al., 1991). Classical biological control of the weevil was planned accordingly. Through cooperation with USDA, APHISPPQ (Bryan et al., 1993), four species of parasitoid wasps, which had been collected in the US, were introduced into Japan in 1988 and 1989; two parasitoids of larvae, Bathyplectes anurus (Thomson) and B. curculionis (Thomson) (Ichneumonidae) and two parasitoids of adults, Microctonus aethiopoides Loan and Microctonus colesi Drea (Braconidae). These parasitoids were cultured in the National Biological Control Laboratory, USDA. The three parasitoid species, except M. colesi, were released in various parts of Kyushu Island from 1989. Usually over 300 (at least 100) adults of one (rarely two or three) species were released per site, and then in the appropriate season for each parasitoid species,
145
follow-up surveys had been held for 2 or 3 years after release except one location, Shiranoe, Moji Ward. However, no parasitoid species were recovered for several years after the release; the establishment appeared to have failed. In 1996, one cocoon of B. anurus was collected unexpectedly in Shiranoe, Moji Ward, Kitakyushu City (33°56⬘N, 131°00⬘E), where the parasitoid had been liberated in 1991 and 1992 but no subsequent surveys had unfortunately been done to conWrm establishment in that place. A cocoon was discovered adjacent to the release point, and there was no other release record of the species in Kitakyushu City or within 20 km around Shiranoe. This suggested that one of introduced parasitoid species became established in Moji Ward. Bathyplectes anurus is a solitary endoparasitoid of larvae of the alfalfa weevil. It has one generation in each year, and attacks the weevil larvae in spring (Bartell and Pass, 1980; Brunson and Coles, 1968; Dowell and Horn, 1977). In the US, B. anurus is one of the key biological control agents of alfalfa weevil. This parasitoid was introduced from Europe and is well established (Berberet and Bisges, 1998; Giles et al., 1994; Harcourt, 1990; OloumiSadeghi et al., 1993; Parr et al., 1993). France was the primary source of the introduction (Dysart and Day, 1976). There has been no study about the establishment, incidence, and eVectiveness of this parasitoid since its introduction in Japan. In 1998, 2 years after the Wrst recovery of B. anurus in Japan, we began a survey of alfalfa weevil and the parasitoid in Kitakyushu City. Our objectives were to conWrm the establishment of B. anurus, and to assess its current eVectiveness in suppressing the alfalfa weevil. Abundance of weevil larvae, extent of damage of the weevil on Chinese milk vetch, and incidence of the parasitoid were investigated in the Weld. The relation between abundance of the weevil larvae in a given year and percentage of parasitism in the previous year was analyzed.
2. Materials and methods Surveys were conducted every spring from 1998 to 2004. Since preliminary observations had shown that emergence of B. anurus adults reached the peak around early April in Northern Kyushu, Weld surveys were conducted from late April to early May. During this period, most of weevil larvae are in their fourth instars, i.e., the last larval instar, and the oviposition period of B. anurus has been ended. Thus, collecting weevil larvae in that period allows assessing the annual parasitism of the weevil. 2.1. Establishment and disperse of the parasitoid Surveys were carried out to estimate the distribution of the parasitoid in Kitakyushu City (Fig. 1) since
146
M. Shoubu et al. / Biological Control 34 (2005) 144–151
Fig. 1. The distribution and expansion of B. anurus in Kitakyushu City during the years of 2000–2003. Solid circle, locality where B. anurus detected; open circle, locality where B. anurus not detected.
1998–2003. Alfalfa weevil larvae were sampled from legumes and Chinese milk vetch and were reared in the laboratory to determine whether the parasitoid was distributed at each sampling location. Sampling locations were selected arbitrarily. Since many of sampling locations were in disturbed environments, like a levee, edge of a road or vacant land in residential areas, we could not collect samples from the same locations continuously throughout the 6-year-investigation period. In 1998, larvae of the weevil were collected at nine locations within a radius of 5 km from the release point of Shiranoe. In 1999, 43 locations scattered throughout Moji Ward were chosen for the survey. The investigation area was expanded to the whole of Kitakyushu City, including Moji Ward, in 2000–2003; we surveyed 24 locations in 2000, 49 in 2001, 48 in 2002, and 34 in 2003. Sampling locations were mostly 1 km apart from each other. Leguminous forbs were beaten as the collector walked linearly up to 35 m. Dropped larvae were recovered in a white, shallow plastic pan. Each sample was put in a paper bag with a small handful of host plants and transported to the laboratory. In the laboratory, the weevil larvae were reared in paper bags. Rearing room was set to 23 °C with a natural photoperiod. To avoid excessive humidity and occurrence of epizootic, a low RH was maintained by an air conditioner. Fresh foliage of greenhouse-grown alfalfa was given to weevil larvae, when necessary. Old desiccated foliage was not taken
out, but was left in the paper bag to prevent loss of larvae, and to provide a pupation site. Feeding was continued for about 2 weeks. Then paper bags were stored in the rearing room. One month later, cocoons of B. anurus were extracted from plant debris by rubbing dried materials, to ascertain whether the parasitoid occurred at each sampling locations. 2.2. Incidence and eVectiveness of the parasitoid A survey was conducted on Chinese milk vetch in 2000–2004 to determine the abundance of weevil larvae, extent of damage of the weevil, and percentages of parasitism by B. anurus. The investigation area was restricted to Moji Ward, because our preliminary survey had shown that B. anurus was relatively abundant and widely distributed in Moji Ward but this was not the case outside Moji Ward. Six (2000), eight (2001–2003), and seven (2004) paddy Welds in Moji Ward, where Chinese milk vetch had been planted, were selected for the survey; from 2001, eight locations had been surveyed including six Welds surveyed in 2000, and one of them was not surveyed in 2004. A 20sweep sample per Weld was taken with a standard sweep net (30.1 cm in diameter) and collected weevil larvae were placed in a paper bag. In the laboratory, each sample was transferred to a shallow pan and the number of collected larvae was counted. Then, they were returned
M. Shoubu et al. / Biological Control 34 (2005) 144–151
to a paper bag and reared in it until cocoons of the parasitoid were formed. Other rearing procedures were the same as described for the survey. The number of parasitoid cocoons was counted for each sample Weld. The mean number of collected weevil larvae was used as an estimate of an annual abundance of weevil larvae on Chinese milk vetch. The percentages of parasitism by B. anurus were assessed for each Weld based on the total numbers of weevil larvae and parasitoid cocoons. The extent of weevil damage on Chinese milk vetch was represented by an index developed by Yamaguchi et al. (1992). With this index, the degree of weevil damage can be expressed between 0 and 100, with larger counts indicating more serious weevil damage; when the count is 0, no leaves of Chinese milk vetch in a Weld are infested; when the count is 100, almost all leaves are heavily infested, and no Xower or few Xorets present. Data were analyzed with the aid of JMP (SAS Institute, 2001). Prior to analysis, data were examined for normality and homogeneity, and if appropriate, parametric procedures, i.e., ANOVA and regression analyses, were applied. Wilcoxon’s tests were used as a non-parametric procedure. The percentages of infestation level and parasitism were transformed to arcsine square roots to stabilize error variances, before being subjected to the analysis. Untransformed means were reported.
3. Results 3.1. Establishment and dispersal of the parasitoid Surveys in 1998–1999 revealed that B. anurus was distributed throughout most of Moji Ward; cocoons of the parasitoid were recovered at 8 points out of 9 in 1998 and 33 points out of 43 in 1999. Therefore, in addition to Moji Ward, leguminous forbs in other wards of Kitakyushu City were sampled for H. postica larvae in 2000–2003. Fig. 1 shows the sampling locations in each investigation year (2000–2003); solid circles indicate locations where at least one parasitoid individual was collected and open circles are locations where no parasitoid was discovered. In 2000, the parasitoid was rarely found except in Moji Ward. Later, the parasitoid was frequently found in areas other than Moji Ward, clearly extending its range of distribution. By 2003, B. anurus was collected from the whole Kitakyushu City (Fig. 1). In 2000, cocoons of parasitoid were recovered from about 50% of samples. The rate has increased gradually; the parasitoid was recovered at all sampling locations in 2003.
147
Fig. 2. The number of weevil larvae collected (A), and the extent of weevil damage on Chinese milk vetch (B), during the years of 2000–2004. The number of weevil larvae and the extent of weevil damage diVered signiWcantly among the years (Wilcoxon’s test, P < 0.01). Data are shown as means § SE.
(Fig. 2A). Infestation by weevil larvae of Chinese milk vetch also varied signiWcantly among the years between 2000 and 2003 (Wilcoxon’s test, df D 4, 2 D 18.26, P D 0.001) (Fig. 2B). The extent of infestation (infestation index) and the weevil numbers reached a peak in 2001 and then declined quickly (Fig. 2). A regression analysis revealed a positive relationship between the extent of infestation and the numbers of weevil larvae (r2 D 0.84, df D 1, F D 110.1, P < 0.0001) (Fig. 3). The results suggested that our sampling method appropriately allowed estimation of the density of weevil larvae.
3.2. Incidence and eVectiveness of the parasitoid The number of weevil larvae collected varied among the years (Wilcoxon’s test, df D 4, 2 D 13.39, P D 0.010)
Fig. 3. The relationship between the extent of weevil damage on Chinese milk vetch and the number of weevil larvae collected in 2002–2004 for each collection site (regression analysis; P < 0.0001).
148
M. Shoubu et al. / Biological Control 34 (2005) 144–151
Fig. 4. The percentages of weevil parasitism by B. anurus on Chinese milk vetch in 2000–2004. The percentages diVered signiWcantly among the years (Wilcoxon’s test, P D 0.0007). Data are shown as means § SE.
The percentages of parasitism of weevil larvae varied signiWcantly among the years between 2000 and 2004 (Wilcoxon’s test, df D 4, 2 D 19.32, P D 0.0007) (Fig. 4). In 2000, parasitoid cocoons were not obtained from twothird of Weld samples; the mean percentage of parasitism (§SE) was 1.92 § 1.38. Thus, B. anurus was rarely found in 2000. However, the percentages of parasitism quickly increased later, and the mean percentage reached 46.56 § 13.43% in 2004 (Fig. 4). In 2003 and 2004, B. anurus was found in all sampling areas. The results evidently showed that the parasitoid had quickly increased its density during the 5 years.
Fig. 5. The relationships between percentages of parasitism in a year and those in the next year (A), and increase in parasitism (in folds) (B) for each collection site (regression analysis; P D 0.004).
To conWrm whether B. anurus was established at each sampling area and whether it increased its population, the relationship between the percentages of parasitism in a year and those in 1 year previously for each area was examined. Because in 2000, B. anurus was rarely collected in many sampling areas, data collected between 2001 and 2004 were used for the analyses. The analyses showed that, for each sampling area, B. anurus parasitism was mostly higher than that 1 year previously (Fig. 5A). Although the percentage of parasitism increased at maximum of 40.62 holds, the degree of parasitism increase depended on the percentage of parasitism 1 year previously (regression analysis after log transformation; r2 D 0.38, df D 1, F D 10.80, P D 0.004) (Fig. 5B). Lastly, it was assessed whether increase in the percentage of B. anurus parasitism could cause reduction of weevil population and hence reduction of damage of Chinese milk vetch. To do this, for each sampling place, the percentages of parasitism in a year were plotted against the numbers of weevil larvae collected in the next year. Regression analysis showed a signiWcant, negative relationship between the two (r2 D 0.50, df D 1, F D 19.7, P D 0.0003) (Fig. 6A). Similarly, a negative relationship between the percentage of parasitism and the degree of
Fig. 6. The relationships between the percentages of parasitism in a year and the number of weevil larvae collected (A), and the extent of weevil damage (B) in the next year for each collection site. Negative relationships were detected for both cases (regression analyses; P < 0.001).
M. Shoubu et al. / Biological Control 34 (2005) 144–151
weevil infestation was detected (r2 D 0.38, df D 1, F D 13.1, P D 0.0016) (Fig. 6B). These results strongly suggested that higher levels of parasitism in a year resulted in lower levels of weevil population and infestation in the next year, demonstrating that biological control worked against alfalfa weevil.
4. Discussion Our survey over 7 years has conWrmed establishment of B. anurus in Japan. In addition, the parasitism rate by B. anurus reached over 40% in 2003 and 2004, and alfalfa weevil populations in Moji Ward declined in response to an increase in parasitism rate. Classical biological control appears to be working here. Classical biological control is often eVective against exotic pests (Caltagirone, 1981). Suppression of alfalfa weevil populations with introduced parasitoids in the US is one of the most successful examples in classical biological control (Day, 1981). Among a number of parasitoid species released in the US, B. anurus became the most prevalent and eVective larval parasitoid, especially in the Northeastern states (Day, 1981) and the central states, like Illinois (Oloumi-Sadeghi et al., 1993) or Iowa (Giles et al., 1994). In those regions, alfalfa weevil population density is usually below the level that causes economic damage due to the contribution of natural enemies including B. anurus (RadcliVe and Flanders, 1998). There are a number of reasons why B. anurus is an eVective agent. Its jumping behavior may increase survival of larvae within cocoons by eluding hyperparasitoids or unfavorable environmental conditions (Day, 1970). Also, B. anurus is not encapsulated by any strain of the host (Maund and Hsiao, 1991; Puttler, 1967). Besides, it has some relatively superior attributes as a natural enemy; great fecundity, rapid searching and handling of hosts, and aggressiveness in eliminating supernumeraries during the larval stage (Bartell and Pass, 1980). Among the other three parasitoid species introduced in Japan, M. colesi could not be released because the mass-rearing had failed. The remaining two parasitoids, B. curculionis and M. aethiopoides, were released several times in various parts of Kyushu Island. In spite of intensive follow-up surveys, these species have not been discovered, and their establishment appears to be unsuccessful. The reasons for failure may include the release of too few individuals, or poor adaptation of the parasitoids to environmental conditions at the release location. Bathyplectes curculionis was the most expected species among parasitoids introduced into Japan because this parasitoid had been shown to establish and spread quickly after the release in some areas of the US, like the Great Basin area (Chamberlin, 1926) and Northeast
149
(Dysart and Day, 1976). In Japan, over 8000 adults in total had been released in several localities in Kyushu Island from 1989 to 1999, but it has not been recovered from the Weld yet. There are some possible reasons for the failure of the establishment. In Northern Kyushu, emergence of B. curculionis does not synchronize with the peak occurrence of the suitable host, i.e., Wrst or second instars of the weevil, whereas B. anurus emergence falls within this period. Because we mostly did not release multiple parasitoid species in each location, interspeciWc competition among parasitoids is unlikely to be a factor causing unsuccessful establishment of the parasitoids. It is well known that the egg of B. curculionis is encapsulated by Egyptian and eastern strains of the weevil (Maund and Hsiao, 1991). EVectiveness of B. curculionis is often comprised seriously by this host defensive reaction (Puttler, 1967; van den Bosch and Dietrick, 1959). Though the weevil strain distributed in Japan has not been clariWed yet, Kimura and Ito (1992) observed that B. curculionis eggs were frequently encapsulated in the body of weevil larvae, which originated from Kitakyushu City (when single parasitoid egg was deposited, the percentage of complete encapsulation was 94.2%, for example). Kuwata et al. (in press) sequenced partial mitochondrial DNA of alfalfa weevil individuals collected from various localities in Japan, and conWrmed that at least two strains of alfalfa weevil invaded Japan, including Kyushu. Egyptian strain can be one of the two strains (Kenji Ohto, unpublished data; Kuwata et al., in press). The weevil defensive reaction may be a factor that obstructed the establishment of B. curculionis in Japan. Only B. anurus became established in Northern Kyushu and appears to be reducing the population density of the alfalfa weevil. Although B. anurus is an eVective agent in classical biological control of alfalfa weevil, our survey suggests that many years may be necessary. The parasitoid was slow to spread in North America. In Northeastern United States, despite of the extensive release during the Wrst national alfalfa weevil biological control project in 1960s, the distribution of B. anurus was restricted around release areas at least by the end of the decade (Dysart and Day, 1976). The parasitoid was also introduced into Ontario, Canada, in 1970, but it was not detected outside of the release area until 1981 (Harcourt, 1990). This appears to be the case in Japan; our survey has revealed that B. anurus is expanding its distribution very slowly. The parasitoid was collected only around the release area even more than a decade after the initial release. The current distribution of the parasitoid does not exceed 40 km from the release point (Okumura et al., unpublished data). Moreover, even when the parasitoid has been once established, the percentage of parasitism is not suYciently high at least for several years. It was shown in Ontario that the parasitism rate by B. anurus was low for
150
M. Shoubu et al. / Biological Control 34 (2005) 144–151
the Wrst several years after parasitoid release, but then rose suddenly (Harcourt, 1990). The parasitoid was established in Oklahoma in 1972. In addition to little tendency for dispersal, however, rates of parasitism had not exceeded 2% for 6 years since establishment (Berberet et al., 1978). It took more than 10 years for the parasitoid to increase its incidence (Berberet and Bisges, 1998). Similarly, our survey has shown that the percentage parasitism by B. anurus on Chinese milk vetch remained at low level (
