thropod predators such as carabid and staphylinid beetles and spiders can efciently reduce the popu- lations of some pest insects and slugs (Best and Beegle, 1977; Edwards et al., 1979; Chiverton, 1986; Clarke et al., 1994; Kromp, 1999). Thus, emphasis has recently been placed on studies to create and manage farmland habitats for augmenta- tion of such benecial predators (Thomas et al., 1991; Lys and Nentwig, 1992; Lys et al., 1994; Kromp, 1999). These studies indicate that eld margins and semi-natural habitats with dense vege- tation harbor high abundance and species richness of benecial predators. These habitats not only support feeding and reproductive activity but also function as overwintering refuges for the predators (Desender, 1982; Sotherton, 1984, 1985; Thomas and Marshall, 1999; Pffner and Luka, 2000). Some studies have shown that predatory epigaeic insects which had overwintered in eld margins emigrated into adjacent farmland in spring (Coombes and Sotherton, 1986; Thomas et al., 1991, 2002). However, these studies were restricted to cereal and forage crop elds in Europe and North America (Kromp, 1999). In Japan and other temperate Asian regions, rice, vegetables and fruits are intensively cultivated on narrow plains and foothill areas of mountains. Farmland management practices in Japan are dif- ferent from those in Europe and North America. Assemblages of carabids and other predators in Japan thus may be distinct from those in other world areas, although Luff (2002) compared the carabid assemblages of various world farmlands in- cluding Japan mainly at the genus level and dis- cussed how dominant genera were common among Appl. Entomol. Zool. 38 (4): 449459 (2003) 449 Ground beetles (Coleoptera: Carabidae) and other insect predators overwintering in arable and fallow elds in central Japan Kazuo YAMAZAKI, 1, * Shinji SUGIURA 2 and Koji KAWAMURA 2 1 Osaka City Institute of Public Health and Environmental Sciences; Tennoji, Osaka 5430026, Japan 2 Laboratory of Forest Ecology, Graduate School of Agriculture, Kyoto University; Kyoto 6068502, Japan (Received 5 February 2003; Accepted 26 May 2003) Abstract To clarify assemblage patterns of overwintering ground beetles (Coleoptera: Carabidae) and other insect predators in farmland habitats for the purpose of proper land management to enhance benecial predators, we collected carabid and other insect predators at eight plots including arable and fallow rice and vegetable elds and a bank of an adjacent irrigation pond in central Japan. In total, 159 adults and 268 larvae of 33 carabid species, and 178 individuals of at least 17 species of other insect predators were collected by the quadrat sampling method. In rice elds, both the num- ber of species and no. of individuals of overwintering carabid beetles increased as the soil became dry and vegeta- tional succession proceeded, whereas in fallow vegetable elds carabids decreased according to succession. Similar trends were conrmed in other insect predators. Variations of carabid species richness and abundances among the plots might be attributed to soil water content, vegetation and prey availability. In early-successional fallow vegetable elds, the larvae of the carabid genus Harpalus overwintered with high density; this appeared to be because the n- gergrass Digitaria ciliaris (Poaceae), whose seeds were a potential food for Harpalus, were densely vegetated there. In a dry fallow rice eld and early-successional vegetable elds, benecial predators such as Dolichus halensis (Coleoptera: Carabidae), Agrypnus binodulus (Coleoptera: Elateridae), and soldier beetle (Coleoptera: Cantharidae) larvae hibernated with high densities. For proper farmland management to augment insect predators, it is desirable to maintain fallow rice and vegetable elds as relatively dry habitats and at early successional stages. Ploughing fallow elds in winter may reduce overwintering predacious insect larvae. Key words: Carabidae; hibernation; benecial predators; pest control; habitat preference * To whom correspondence should be addressed at: E-mail: K.Yamazaki@iphes.city.osaka.jp temperate world farmlands. Therefore, in order to use carabid beetles and other benecial predators as biological control agents in Japan, the assem- blage patterns in each habitat and habitat prefer- ences of component species should be claried. In Japan, Habu and Sadanaga (1961, 1963, 1965, 1969, 1970a, b, 1971) described larval mor- phology of carabid beetles and noted their biology in cultivated elds and paddy elds based on labo- ratory rearings. Carabid assemblages and faunae have been described from paddy elds, forage crop eld, cabbage eld, g orchards and vineyards dur- ing the warm season using pitfall-trapping (Torii, 1974; Yano et al., 1989, 1995; Yahiro et al., 1992; Ishitani et al., 1994; Ishitani and Yano, 1994; Ishi- tani, 1996; Suenaga and Hamamura, 2001; Yano, 2002). Yahiro and Yano (1997) reported long-term data on a carabid assemblage caught by a light trap set at a farmland area which comprised rice, cereal and vegetable elds, orchards and irrigation ponds. However, in Asia including Japan, no quantitative study on overwintering carabid assemblages in farmland and adjacent semi-natural habitats has been reported. Quantitative assemblage data of hi- bernating carabids and other insect predators in each farmland habitat is useful to create and man- age overwintering refuges for the predators. Carabid life-cycle patterns (breeding season and hibernating developmental stage) vary across cli- matic regions and habitats (Murdoch, 1967; Paar- mann, 1979; Andersen, 1984). However, the life cycles of most carabid species excluding the tribe Carabina are not well known in Japan (for Cara- bina see Sota, 1985). Collecting ground beetles in winter helps to elucidate their life cycles, espe- cially their hibernating stages. Information on life cycles of benecial carabid species may be essen- tial for proper land-management practices to aug- ment those already in use; In Europe, deep plough- ing during autumn or winter is known to reduce the numbers of overwintering carabid beetles and other predators (Kromp, 1999; Pffner and Luka, 2000). This negative effect to carabids may be severe for hibernating larvae compared with overwintering adults, since the larvae appear to be more vulnera- ble to physical disturbance than the adults. In the present study, we sampled overwintering carabid beetles and other insect predators in arable and fallow elds (rice and vegetables) with differ- ent successional stages and a bank of adjacent irri- gation ponds in central Japan. Firstly, assemblages of carabids and other predators in each habitat are described, and the habitats with high abundance and species richness of the predators are explored. Similarities between assemblages are also ana- lyzed. Overwintering developmental stages and breeding seasons of the sampled carabid beetles are compared among habitats. Then, habitat prefer- ences for benecial species are examined, and in turn land management practices to enhance bene- cial insect predators against pest insects in Japa- nese agroecosystems are discussed. MATERIALS AND METHODS Study site. The study site was located in a rural area of Son-enji (3448N, 13543E, ca. 100 m above sea level), Hirakata City, Osaka Prefecture, central Japan (Fig. 1). The landscape of this area is well preserved, comprising mosaics of coppice woodlands, rice and vegetable elds, irrigation ponds and streams, and therefore relatively rare plants and insects that prefer grassland and marsh- land habitats were still extant (Kanky-kagaku Co., 2001). There were many arable and fallow rice and vegetable elds at the study site. Fallow elds were at various successional stages. These farmlands were surrounded by secondary forests composed mainly of the Japanese red pine Pinus densiora and the oak Quercus serrata. Overwintering ground beetles and other insect predators were sampled at the following eight plots which were within a 1 km 2 area. Table 1 shows the environmental conditions of sampling plots. Age after cultivation in each plot was estimated by con- sulting with farmers and based on vegetational suc- cession. Vegetational succession advanced from A to C in rice elds, and from D to G in vegetable elds. Methods. We arbitrarily set three quadrats (1.6 m1.6 m) per plot on the ground except the two plots (D: 2 quadrats and F: 1 quadrat), that were unfortunately ploughed while sampling. All the quadrats except the bank of an irrigation pond were set on relatively at and horizontal ground. Ground inclination of three quadrats on the pond bank was ca. 50. We then dug up each quadrat to a depth of 40 cm using hoes, and inspected the soil thus obtained for ground beetles and other insect predators (see Yamazaki et al., 1999, 2002). The 450 K. YAMAZAKI et al. sampling was conducted between December 1999 and March 2000 and in March 2001. As to ants, only queens were caught because of the difculty and the labor intensity required for sampling work- ers. In the present study, we focused on epigaeic and soil macro-predators, but did not examine predators on plant surfaces such as lacewings and syrphids and other micro-predators. The collected carabid larvae were identied according to the de- scriptions of Habu and Sadanaga (1961, 1963, 1965, 1969, 1970a, b, 1971) and our collections (Yamazaki et al., 1999, 2002). The abundance and species richness of carabid beetles and other predators per quadrat were com- pared among the sampling plots. A one-way ANOVA model was used to compare the number of species and abundance among the plots in carabid beetles and other predators, and Scheffs range test was also used to clarify the difference. Plots D and F were excluded from these analyses due to the lack of an adequate data set. The similarities among assemblages were compared using Wards minimum-variance clustering method. Prior to the one-way ANOVA and cluster analysis, the no. of individuals was logarithmically transformed. Over- wintering developmental stages and breeding sea- sons in carabids were compared among the plots. Breeding seasons were determined on the basis of the descriptions by Kurosa (1959), Habu and Sadanaga (1961, 1963, 1965, 1969, 1970a, b, 1971), Sota (1985) and Ishitani (1996). Chi-square test was used to ascertain whether the percentage (no. of individuals) of larvae and autumn-breeders differed among the plots. Overwintering Ground Beetles in Rice and Vegetable Fields 451 Fig. 1. Study site of the ground beetles and other insect predators overwintering in farmland habitats. A to H denote sampling plots. Table 1. The environmental conditions of sampling plots Plots Land use Years after Dominant plants abandonment A Arable rice eld Astragalus sinicus (Leguminosae), grasses B Wet fallow rice eld 1 Oenanthe javanica (Umbelliferae), Stellata sp. (Caryophyllaceae) C Dry fallow rice eld 1 Erigeron canadensis (Asteraceae), A. sinicus D Fallow vegetable eld 1 0.5 Digitaria ciliaris (Poaceae), E. canadensis E Fallow potato eld 1 D. ciliaris F Fallow vegetable eld 2 1 Solidago altissima, Artemisia indica var. maximowiczii (Ateraceae) G Old fallow eld 3 Miscanthus sinensis, Pleioblastus fortunei (Poaceae) H Bank of irrigation pond Mostly bare ground RESULTS Overview of predator assemblages In a total of eight plots, 605 individuals of cara- bid beetles and other insect predators belonging to at least 50 species were collected by the quadrat sampling method. Among them, carabid beetles predominated in the assemblage, with 427 individ- uals (70.6% of all the predators) belonging to 33 species (66.0% of all the predator species) that were collected. The collected carabid beetles con- sisted of 159 adults (37.2%) of 27 species and 268 larvae (62.8%) of nine species (Table 2). In addi- tion, one adult and 63 larvae (36.0% of all other predators) of four click beetle species (Elateridae) were sampled, and 47 larvae (26.4%) of several soldier beetle species (Cantharidae) and 26 adults and four larvae (16.9%) of ve rove beetle species 452 K. YAMAZAKI et al. Table 2. Collected carabid beetles overwintering at eight plots of arable and fallow elds. Numbers without and in parentheses are adults and larvae, respectively. denotes none found. Plot Carabid species Breeding Rice elds Vegetable elds Pond Total season a A B C D E F G H Carabus yaconinus yaconinus Bates S 1 1 Leptocarabus kumagaii Komiya et Kimura A (5) (5) Scarites terricola pacicus Bates S 3 3 Lesticus magnus (Motschulsky) S 1 1 Pterostichus sulcitarsis Morawitz S 2 2 P. haptoderoides japanensis Lutshnik A 10 3 13 P. longinquus Bates S 1 23 1 25 P. microcephalus (Motschulsky) S 2 2 Agonum chalcomus (Bates) ? 2 3 3 8 Colpodes japonicus (Motschulsky) ? 1 1 Dolichus halensis (Schaller) A (1) (6) (8) (1) (16) Amara congrua Morawitz S 15 3 18 Am. chalcites Dejean S 12 1 13 Anisodactylus signatus (Panzer) S 1 1 2 13 17 An. puctatipennis Morawitz S 1 1 ?Harpalus vicarius Harold A (18) (18) H. jureceki (Jedlicka) A (3) (2) 2 1 3(5) H. griseus (Panzer) A (5) (2) (7) H. tridens Morawitz A (13) (13) (26) H. sinicus Hope A (75) (14) 3 3(89) H. niigatanus Schauberger A 1(2) (7) (92) 1(101) H. chalcentus Bates S 5 8 13 Oxycentrus argutoroides (Bates) ? 3 3 Trichotichnus congruus (Motschulsky) ? 1 1 T. noctuabundas Habu ? 1 1 Bradycellus subditus (Lewis) ? 13 13 B. grandiceps (Bates) ? 6 6 Acupalpus inornatus Bates S 1 1 Stenolophus iridicolor Redtenbacher S 5 5 S. fulvicornis Bates S 1 1 Anoplogenius cyanescens (Hope) S 2 2 Chlaenius naeviger Morawitz S 1 1 Carabidae Gen. sp. ? (1) (1) Total 4 33(1) 63(8) 6(104) 14(134) 27(21) 5 7 159(268) a S: spring, A: autumn. Breeding seasons of each species were determined based on Kurosa (1959), Habu and Sadanaga (19611971), Sota (1985) and Ishitani (1996). (Staphylinidae) were also collected (Table 3). Assemblage patterns according to vegetational succession The number of species and the no. of individuals of carabids per quadrat were signicantly different among the sampling plots (one-way ANOVA, species number, df5, F6.24, p0.01; no. of in- dividuals, df5, F17.73, p0.0001) (Fig. 2). In rice elds (plots AC), both the no. of species and individuals of carabids increased, from arable A to C (one year old elds), as the soil became dry and succession proceeded. On the other hand, in fallow vegetable elds (plots DG), there was a tendency for overwintering carabid beetles to decrease as succession proceeded: The no. of species was simi- lar among plots DF, but plot G harbored a small number of carabid species (Fig. 2a). The no. of in- dividuals decreased according to vegetational suc- cession (Fig. 2b). The number of species and the no. of individuals of predators other than carabids per quadrat were also signicantly different among the sampling plots (one-way ANOVA, species number, df5, F4.96, p0.05; No. of individuals, df5, F7.20, p0.01) (Fig. 3). In rice elds, both Overwintering Ground Beetles in Rice and Vegetable Fields 453 Table 3. Collected insect predators except Carabidae at eight plots of arable and fallow elds. Numbers in and without parentheses are adults and larvae, respectively. denotes none found. Plot Predator species Rice elds Vegetable elds Pond Total A B C D E F G H Histeridae Atholus pirithous (Marseul) 1 1 Staphylinidae Olophrum arrowi Scheerpeltz 1 1 Paederus fuscipes (Curtis) 17 17 Lathrobium unicolor Kraatz 1 1 Tachyporus celatus Sharp 1 1 1 3 6 Aleochara sp. 1 1 Staphylinidae Gen. sp. (2) (1) (1) (4) Elateridae Agrypnus binodulus (Motschulsky) (23) (35) (58) Ampedus sp. (2) (2) Melanotus cete Candeze 1 1 Melanotus sp. (1) (2) (3) Cantharidae Cantharidae Gen. spp. (1) (7) (6) (28) (4) (1) (47) Coccinellidae Coccinella septempunctata L. 3 3 Formicidae Crematogaster osakensis Forel a 10 5 15 Vespidae Vespa tropica pulchra Buysson a 2 2 Reduviidae Peirates turpis Walker 1 1 Gryllotalpidae Gryllotalpa orientalis Burmeister 1 (2) (5) 1 2(7) Anisolabididae Euborellia plebeja (Dohrn) 6 6 Total 1(1) 17(9) 22(36) 3 6(64) 5(5) 1(6) 2 57(121) No. of quadrats 3 3 3 1 3 2 3 3 21 a Queens. species richness and abundance of the other preda- tors increased with succession, similar to the case of carabid beetles. In vegetable elds, although the abundance of the other predators was high at plot E (fallow potato eld) (Fig. 3b), the no. of species and abundance were similar among the other plots. Among the eight plots, the average number of species of the other predators per quadrat was sig- nicantly correlated to that of carabids (Spear- mans correlation coefcient by rank test, N8, rs0.79, p0.05) (Figs. 2a, 3a), although the aver- age no. of individuals of other predators per quadrat was marginally correlated to that of cara- bids (Spearmans correlation coefcient by rank test, N8, rs0.64, p0.08) (Figs. 2b, 3b). This nding suggested that habitat preferences in cara- bids and other predators were similar. Predator assemblages at eight plots Predator assemblage patterns at each plot are briey described as follows (Tables 2, 3): (A) The arable rice eld harbored very small numbers of insect predators (Figs. 1, 2); (B) Wet fallow rice eld: A hygrophilous carabid, Pterostichus longiquus was the dominant species (23 individu- als, 67.6% of carabid individuals), and Acupalpus inornatus, Stenolophus iridicolor and Paederus fuscipes were found only from this plot; (C) Dry fallow rice eld: Carabid (71 individuals of 12 species), staphylinid (four individuals of three species) and elaterid beetles (23 Agrypnus binodu- lus larvae) were relatively dominant, and Lepto- carabus kumagaii, Bradycellus subditus and the earwig Euborellia plebeja were specic to the plot. Two carabid species of the genus Amara were 454 K. YAMAZAKI et al. Fig. 2. Carabid species number and abundance at eight plots of arable and fallow elds. (a) Species number per quadrat, (b) No. of individuals per quadrat. Means1SE. Different alphabetical letters above the bars indicate statistical difference at p0.05 level of signicance by Scheffs range test. Before analysis, no. of species and individuals were logarithmically transformed. Fig. 3. Species number and abundance of insect predators other than carabid beetles at eight plots of arable and fallow elds. (a) Species number per quadrat, (b) No. of individuals per quadrat. Means1SE. Different alphabetical letters above the bars indi- cate statistical difference at p0.05 level of signicance by Scheffs range test. Before analysis, no. of species and individuals were logarithmically transformed. abundantly collected (27 individuals); (D) Fallow vegetable eld 1: This plot was characterized by the high abundance of Harpalus larvae (98 individ- uals of four species). (E) Fallow potato eld: Cara- bid (148 individuals of eight species), elaterid (36 individuals of two species) and cantharid beetles (28 individuals) were abundant. Larvae of ve Harpalus species occurred with high density simi- lar to plot D; (F) Fallow vegetable eld 2: The carabids Scarites terricola pacicus, P. sulcitarsis and B. grandiceps were found only in this plot; (G) Old fallow eld: The abundance of predators was very low (Figs. 2, 3), but two species of the carabid genus Stenolophus were only in this plot; (H) Bank of irrigation pond: Both species richness and abun- dance of predators were very low (Figs. 2, 3). Sil- vicolous species, Carabus yaconinus yaconinus, Chlaenius naeviger and the hornet Vespa tropica pulchra were sampled. Similarities among assemblages Dendrograms of Fig. 4 show the similarities among the assemblages of the plots in carabids and other predators, respectively. For carabid assem- blages, the fallow eld plots D and E were similar, but three rice eld plots (AC), especially the dry fallow rice eld (C), were distinct from each other (Fig. 4a). For other predator assemblages, the arable rice eld (A) and the bank of the pond (H) were similar, possibly due to the small number of predators sampled (Figs. 2, 3). The dry fallow rice eld (C) bore a distinct assemblage of insect preda- tors from other plots (Fig. 4a). Overwintering developmental stages and breed- ing seasons in sampled carabids Table 4 shows the overwintering developmental stages and breeding seasons in the collected cara- bid beetles at each plot. In rice elds (AC), there was a tendency for the percentage of the no. of species and individuals of larvae that overwintered to increase somewhat from arable (A) to fallow (B) and (C). In vegetable elds (DG), as to both no. of species and individuals however, the percentage of larvae that hibernated decreased markedly and the percentage of adults increased with vegeta- tional succession. Plots D and E had many over- wintering Harpalus larvae (cf. Table 2). From the pond bank (H), a small number of adults was found. With regard to breeding seasons, in the dry fal- low rice eld (C), autumn-breeders were found, while in both the arable rice eld (A) and wet fal- low rice eld (B), no autumn-breeders were col- lected. In fallow vegetable elds (DG), with re- gard to both the no. of species and individuals, the percentage of autumn-breeders decreased and the percentage of spring-breeders tended to increase as succession proceeded. A small number of adult carabids collected at the pond bank (H) comprised both spring- and autumn-breeders. DISCUSSION Assemblage patterns according to vegetational succession In this study, we could not examine arable veg- etable elds and old rice elds, and our sampling was restricted to only eight plots. Therefore, the ndings presented in this paper are snapshots of overwintering predators in farmland. The present results may, however, reect some important as- semblage patterns during vegetational succession in farmlands. In rice elds (from A to C), overwin- Overwintering Ground Beetles in Rice and Vegetable Fields 455 Fig. 4. Dendrograms showing the assemblage similarities of (a) carabid and (b) other insect predators hibernated at eight plots in arable and fallow elds. Wards minimum-variance clustering method was applied. Prior to clustering, no. of individuals was logarithmically transformed. AC: rice elds, DG: vegetable elds, H: pond bank. tering carabid beetles increased as the soil became dry and vegetational succession proceeded, while in fallow vegetable elds (from D to G) carabids decreased according to succession (Fig. 2). These trends also applied more or less to other insect predators (Fig. 3). The increase in carabids in the dry fallow rice eld may be attributed to the loss of water content in the soil and vegetational develop- ment. In the dry fallow rice eld, the carabids P. haptoderoides japanensis, two Amara species and B. subditus were abundantly found (Table 2). Since these carabid species are known to inhabit rela- tively dry elds and orchards (Habu and Sadanaga, 1961, 1963; Ishitani and Yano, 1994; Ishitani, 1996), these xerophilous species appeared to over- winter where they were active during the warm season. Moreover, the low water content in the soil may be advantageous for carabids to hibernate. Frmbs (1994) reported that even hygrophilous species living in peat bogs migrated to drier hum- mocks for overwintering. In Europe, high densities of overwintering carabids and other predators have been found in dense vegetation (Sotherton, 1984; Thomas et al., 1991; Pffner and Luka, 2000). Dense vegetation such as tussock-forming grasses provides less variable temperature environments (Bossenbroek et al., 1977). Such temperature buffering properties of dense vegetation may favor overwintering carabids. In addition, vegetation fos- ters many prey species for carabid predators. Thomas et al. (1992) found that a carabid, Deme- trias atricapillus fed on prey even in winter, and that D. atricapillus abundantly overwintered at dense vegetation areas, but that no signicant rela- tionships were found between the no. of D. atri- capillus and their potential prey among habitats. In fallow vegetable elds, carabid abundance of larval stage and autumn-breeders decreased from D to G, according to vegetational succession (Table 4). This was because many carabid larvae of the 456 K. YAMAZAKI et al. Table 4. Overwintering developmental stages and breeding seasons of carabid beetles collected at eight plots of arable and fallow elds Plot Rice elds Vegetable elds Pond Total A B C D E F G H Overwintering stages No. of species Adults 3 5 8 3 2 7 4 5 24 Larvae 0 1 2 5 6 3 0 0 6 Adults/larvae 0 0 1 0 0 0 0 0 3 Total 3 6 11 8 8 10 4 5 33 Larvae (%) 0.0 16.7 18.2 62.5 75.0 30.0 0.0 0.0 18.2 No. of individuals Adults 4 33 63 6 14 27 5 7 159 Larvae 0 1 8 104 134 21 0 0 268 Total 4 34 71 110 148 48 5 7 427 Larvae (%) a 0.0 ab 2.9 a 11.3 a 94.5 c 90.1 c 43.8 b 0.0 ab 0 ab 62.8 Breeding seasons No. of species Spring 1 4 6 2 1 4 4 2 16 Autumn 0 0 4 6 1 5 0 1 9 Unknown 2 2 2 0 6 1 0 2 8 Autumn (%) 0.0 0.0 33.3 75.0 12.5 50.0 0.0 20.0 27.3 No. of individuals Spring 1 30 36 3 13 16 5 2 106 Autumn 0 0 19 107 134 26 0 1 287 Unknown 3 4 16 0 1 6 0 4 34 Total 4 34 71 110 148 48 5 7 427 Autumn (%) b 0.0 abc 0.0 a 26.8 b 97.3 d 90.5 d 54.2 c 0.0 abc 14.3 abc 67.2 a The percentages of larval stage differed among plots (chi-square test, df7, c 2 263.4, p0.0001). b The percentages of autumn breeders differed among plots (chi-square test, df7, c 2 23.51, p0.0001). a, b The values with different alphabetical letters are signicantly different (sequential Bonferroni test (Rice, 1989), a0.05). genus Harpalus were found from the one-year-old fallow elds D and E but no Harpalus was found from the more than three-year-old fallow eld G (Table 2). All the sampled Harpalus species except H. chalcentus are known as autumn-breeders and overwinter as adults and larvae (Habu and Sadanaga, 1961, 1965, 1970a). Many Harpalus species larvae feed on the seeds of Poaceae, Polyg- onaceae, Asteraceae and so on (Kirk, 1972; Jor- gensen and Toft, 1997; Kromp, 1999; Tooley and Brust, 2002). Since at plots D and E, D. ciliaris densely covered the ground, the Harpalus larvae sampled appeared to feed on the seeds of D. cil- iaris and hibernated at the sites. Yamazaki et al. (1999) reported that Harpalus larvae (especially H. capito and H. sinicus) overwintered with high den- sities in marshy ground along a river where grasses vegetated. As vegetational succession proceeded, D. ciliaris and other grasses declined, possibly re- sulting in the loss of Harpalus larvae in plot G. Luff (2002) compared dominant carabid genera among world farmlands and stated that Harpalus species were particularly predominant in the Japa- nese agricultural fauna. Habitat preferences in benecial species and farmland management Although these Harpalus larvae are not preda- tors but seed-feeders of weeds, some of those adults have omnivorous feeding habits, sometimes feeding on insects (Ishitani, 1996). Furthermore, certain Harpalus species can be used for biological control of injurious weeds (Brust, 1994; Kromp, 1999; Tooley and Brust, 2002). Therefore, aug- mentation of Harpalus species in fallow elds ap- pears to be benecial for farmland management. Ploughing fallow elds in winter may reduce the populations of these Harpalus larvae, since larval stage appears to be more vulnerable than adult stage due to the fragility of larval integuments. Dolichus halensis is a benecial predacious carabid attacking larvae of the diamondback moth Plutella xylostella and other lepidopteran and planthopper pests (Habu and Sadanaga, 1963; Wakisaka et al., 1991), and in the present study the plots D and E harbored this species with relatively high densities (Table 2). This carabid is an autumn- breeder and overwinters as larvae, so ploughing fallow elds in winter may negatively affect this species. Among other insect predators, Agrypnus binodu- lus (Elateridae) larvae and cantharid larvae were abundantly found, especially in the dry fallow rice eld C and fallow potato eld E. A. binodulus larva is known to be predacious (Kurosa, 1959; hira, 1968), but its feeding preferences are not known. In Europe, cantharid larvae are voracious predators feeding on earthworms, aphids and other inverte- brates (Langenstck et al., 1998) and have high dispersal ability in farmlands (Traugott, 2002). In Japan, however, although the taxonomy of adult cantharid species has been progressively devel- oped, its larval biology remains little known (Imasaka, 1998). Bionomics of these elaterid and cantharid species should be claried to use these potentially benecial predators for pest control. In conclusion, the ndings of the present study suggest that, in farmland management, in order to increase the numbers of many carabids and other predators, the rice elds and vegetable elds should be maintained in a considerably dry state (but, to sustain regional biodiversity, some rice elds should maintain permanent waters; e.g. Hi- daka, 1998). Furthermore, because the old (more than three-year-old) fallow elds harbored only a small number of carabid and other predators, it is recommended to maintain fallow elds at an early successional stage. In addition, ploughing fallow elds in winter may reduce the predators, espe- cially carabid, elaterid and cantharid larvae, since larvae appear to be more vulnerable to physical disturbances than adult stages. To validate these farmland management practices in Japan, further eld data on predatory insects overwintering in farmlands of various districts and experimental studies that manipulate management practices are required. ACKNOWLEDGEMENTS We thank the farmers for permitting us to conduct the pres- ent study and two anonymous reviewers for improving the manuscript. Thanks are also due to Dr. Y. Hayashi (Kawanishi City, Hyogo Prefecture) for identifying staphilinid beetles. REFERENCES Andersen, J. (1984) A re-analysis of the relationship between life cycle patterns and the geographical distribution of Fennoscandian carabid beetles. J. Biogeogr. 11: 479 489. Best, R. L. and C. C. Beegle (1977) Food preferences of ve species of carabids commonly found in Iowa cornelds. Overwintering Ground Beetles in Rice and Vegetable Fields 457 Environ. Entomol. 6: 912. Bossenbroek, P. H., A. Kessler, A. S. N. Liem and L. Vlijm (1977) The signicance of plant growth-forms as shel- ter for terrestrial animals. J. Zool. 182: 16. Brust, G. E. (1994) Seed predators reduce broadleaf weed growth and competitive ability. Agric. Ecosyst. Environ. 48: 2734. Chiverton, P. A. (1986) Predatory density manipulation and its effect on the populations of Rhopalosiphum padi (Hom. Aphididae) in spring barley. Ann. Appl. Biol. 109: 4960. Clarke, M. S., J. M. Luna, N. D. Stone and R. R. Youngman (1994) Generalist predator consumption of armyworm (Lepidoptera: Noctuidae) and effect of predator removal on damage in no-till corn. Environ. Entomol. 23: 617622. Coombes, D. S. and N. W. Sotherton (1986) The dispersal and distribution of polyphagous predatory Coleoptera in cereals. Ann. Appl. Biol. 108: 461474. Desender, K. (1982) Ecological and faunal studies on Coleoptera in agricultural land II. Hibernation of Cara- bidae in agro-ecosystems. Pedobiologia 23: 295303. Edwards, C. A., K. D. Sunderland and K. S. George (1979) Studies on polyphagous predators of cereal aphids. J. Appl. Ecol. 16: 811823. Frmbs, H. (1994) The importance of habitat structure and food supply for carabid beetles (Coleoptera, Carabidae) in peat bogs. Mem. Entomol. Soc. Can. 169: 145159. Habu, A. and K. Sadanaga (1961) Illustrations for identica- tion of larvae of the Carabidae found in cultivated elds and paddy-elds (I). Bull. Nat. Inst. Agr. Sci., Tokyo (C) 13: 207248 (in Japanese with English summary). Habu, A. and K. Sadanaga (1963) Illustrations for identica- tion of larvae of the Carabidae found in cultivated elds and paddy-elds (II). Bull. Nat. Inst. Agr. Sci., Tokyo (C) 16: 151179 (in Japanese with English summary). Habu, A. and K. Sadanaga (1965) Illustrations for identica- tion of larvae of the Carabidae found in cultivated elds and paddy-elds (III). Bull. Nat. Inst. Agr. Sci., Tokyo (C) 19: 81216 (in Japanese with English summary). Habu, A. and K. Sadanaga (1969) Illustrations for identica- tion of larvae of the Carabidae found in cultivated elds and paddy-elds (Supplement I). Bull. Nat. Inst. Agr. Sci., Tokyo (C) 23: 113144 (in Japanese with English summary). Habu, A. and K. Sadanaga (1970a) Descriptions of some lar- vae of the Carabidae found in cultivated elds and paddy- elds (I). Konty 38: 923 (in Japanese with English summary). Habu, A. and K. Sadanaga (1970b) Descriptions of some lar- vae of the Carabidae found in cultivated elds and paddy- elds (II). Konty 38: 2441 (in Japanese with English summary). Habu, A. and K. Sadanaga (1971) Descriptions of some lar- vae of the Carabidae found in cultivated elds and paddy- elds (III). Konty 39: 156166 (in Japanese with Eng- lish summary). Hidaka, K. (1998) Biodiversity conservation and environ- mentally regenerated farming system in rice paddy elds. Jpn. J. Ecol. 48: 167178 (in Japanese). Imasaka, S. (1998) Whats joh kai bon?On the Japanese Cantharidae. The Nature and Insects 33(11): 712 (in Japanese). Ishitani, M. (1996) Ecological studies on ground beetles (Coleoptera: Carabidae, Brachinidae) as environmental indicators. Misc. Rep. Hiwa Mus. Nat. Hist. 34: 1110 (in Japanese with English summary). Ishitani, M., J. Watanabe and K. Yano (1994) Species com- position and spatial distribution of ground beetles (Coleoptera) in a forage crop eld. Jpn. J. Entomol. 62: 275283. Ishitani, M. and K. Yano (1994) Species composition and seasonal activities of ground beetles (Coleoptera) in a g orchard. Jpn. J. Entomol. 62: 201210. Jorgensen, H. B. and S. Toft (1997) Role of granivory and insectivory in the life cycle of the carabid beetle Amara similata. Ecol. Entomol. 22: 715. Kanky-kagaku Co. (2001) Guidebook of Nature in Osaka. Osaka Green Trust, Osaka. 121 pp. (in Japanese). Kirk, V. M. (1972) Seed-caching by larvae of two ground beetles, Harpalus pensylvanicus and H. erraticus. Ann. Enotmol. Soc. Am. 65: 14261428. Kromp, B. (1999) Carabid beetles in sustainable agriculture: a review on pest control efcacy, cultivation impacts and enhancement. Agric. Ecosyst. Environ. 74: 187228. Kurosa, K. (1959) Carabidae, Agrypnus binodulus Motschul- sky. In Illustrated Insect Larvae of Japan (T. Esaki et al. eds.). Hokuryukan, Tokyo, pp. 399414, 464 (in Japa- nese). Langenstck, C., U. Heimbach and O. Larink (1998) Larvae of Cantharidae (Insecta: Coleoptera) in arable elds in the south-east of Lower Saxony (Germany) and aspects of their biology. Braunschw. Naturkundl. Schr. 5: 551568. Luff, M. L. (2002) Carabid assemblage organization and species composition. In The Agroecology of Carabid Bee- tles (J. M. Holland ed.). Intercept, Andover, pp. 4179. Lys, J.-A. and W. Nentwig (1992) Augmentation of bene- cial arthropods by strip-management 4. Surface activity, movements and activity density of abundant carabid bee- tles in a cereal eld. Oecologia 92: 373382. Lys, J.-A., M. Zimmerman and W. Nentwig (1994) Increase in activity density and species number of carabid beetles in cereals as a result of strip-management. Entomol. Exp. Appl. 73: 19. Murdoch, W. W. (1967) Life history patterns of some British Carabidae (Coleoptera) and their ecological signicance. Oikos 18: 2532. hira, H. (1968) An overview of elaterid larvae in Japan. The Nature and Insects 3(2): 2429 (in Japanese). Paarmann, W. (1979) Ideas about the evolution of various annual reproduction rhythms in carabid beetles of the dif- ferent climatic zones. Misc. Pap., Land. Wag. 18: 119 132. Pffner, L. and H. Luka (2000) Overwintering of arthropods in soils of arable elds and adjacent semi-natural habitats. Agric. Ecosyst. Environ. 78: 215222. Rice, W. R. (1989) Analyzing tables of statistical tests. Evolution 43: 223225. 458 K. YAMAZAKI et al. Sota, T. (1985) Life history patterns of carabid beetles be- longing to the subtribe Carabina (Coleoptera, Carabidae) in the Kinki district, western Japan. Konty 55: 370 378. Sotherton, N. W. (1984) The distribution and abundance of predatory arthropods overwintering on farmland. Ann. Appl. Biol. 105: 423429. Sotherton, N. W. (1985) The distribution and abundance of predatory Coleoptera overwintering in eld boundaries. Ann. Appl. Biol. 106: 1721. Suenaga, H. and T. Hamamura (2001) Occurrence of carabid beetles (Coleoptera: Carabidae) in cabbage elds and their possible impact on lepidopteran pests. Appl. Ento- mol. Zool. 36: 151160. Thomas, C. F. G., J. M. Holland and N. J. Brown (2002) The spatial distribution of carabid beetles in agricultural land- scapes. In The Agroecology of Carabid Beetles (J. M. Holland ed.). Intercept, Andover, pp. 305344. Thomas, C. F. G. and E. J. P. Marshall (1999) Arthropod abundance and diversity in differently vegetated margins of arable elds. Agric. Ecosyst. Environ. 72: 131144. Thomas, M. B., H. J. Mitchell and S. D. Wratten (1992) Abi- otic and biotic factors inuencing the winter distribution of predatory insects. Oecologia 89: 7884. Thomas, M. B., S. D. Wratten and N. W. Sotherton (1991) Creation of island habitats in farmland to manipulate populations of benecial arthropods: predator densities and emigration. J. Appl. Ecol. 28: 906917. Tooley, J. and G. E. Brust (2002) Weed seed predation by carabid beetles. In The Agroecology of Carabid Beetles (J. M. Holland ed.). Intercept, Andover, pp. 215229. Torii, T. (1974) The population dynamics of the carabid bee- tles trapped at the levee of the paddy eld. Mushi 47: 119126. Traugott, M. (2002) Dispersal power, home range and habi- tat preference of cantharid larvae (Coleoptera: Canthari- dae) in arable land. Eur. J. Soil Biol. 38: 7983. Wakisaka, S., R. Tsukuda and F. Nakasuji (1991) Life tables of the diamondback moth, Plutella xylostella (L.) (Lepi- doptera: Yponomeutidae) and effects of rainfall, tempera- ture and host plants on survival and reproduction. Jpn. J. Appl. Entomol. Zool. 35: 115122 (in Japanese with English summary). Yahiro, K., T. Fujimoto, M. Tokuda and K. Yano (1992) Species composition and seasonal abundance of ground beetles (Coleoptera) in paddy elds. Jpn. J. Entomol. 60: 805813. Yahiro, K. and K. Yano (1997) Ground beetles (Coleoptera, Carabidae) caught by a light trap during ten years. Esakia 37: 5769. Yamazaki, K., S. Sugiura and K. Kawamura (1999) Over- wintering ground beetle assemblages along the Kizu River, Kyoto. Entomol. Sci. 2: 3340. Yamazaki, K., S. Sugiura and K. Kawamura (2002) Environ- mental factors affecting the overwintering distribution of ground beetles (Coleoptera: Carabidae) on a forest oor in central Japan. Entomol. Sci. 5: 125130. Yano, K. (2002) Insect Fauna in Paddy Field. Tokai Univ. Press, Tokyo. 175 pp. (in Japanese). Yano, K., M. Ishitani and K. Yahiro (1995) Ground beetles (Coleoptera) recorded from paddy elds of the world: a review. Jpn. J. Syst. Entomol. 1: 105112. Yano, K., K. Yahiro, M. Uwada and T. Hirashima (1989) Species composition and seasonal abundance of ground beetles (Coleoptera) in a vineyard. Bull. Fac. Agr. Yama- guchi Univ. 37: 114. Overwintering Ground Beetles in Rice and Vegetable Fields 459