Thrips palmi (melon thrips)

Thrips palmi was first described by H. Karny in 1925 from specimens collected in 1921 on tobacco in Sumatra. It was named after the director of the Medan quarantine station, Dr BT Palm, and although it has been referred to as 'palm thrips' (CIE, 1986) it is not known to be associated with any palm tree species. One of the first published records of this species as a pest was an account from southern India of damage to sesame seed pods as a result of feeding on young ovary walls (Ananthakrishnan, 1955). In the Philippines, Medina (1980) reported that an outbreak of T. palmi in 1977 had destroyed almost 80% of the watermelon plantations in central Luzon and Laguna.The oldest available voucher specimens, apart from the types from Sumatra, are four females from India collected at Coimbatore in 1929 and preserved in the collections of the Senckenberg Museum, Frankfurt, Germany (Bhatti, 1980). In the collections of the Natural History Museum, London, and derived from the identification service of the CIE, there are voucher specimens of T. palmi from Thailand in 1947, India (West Bengal) in 1950, Pakistan (North West Himalaya) in 1951, Malaysia in 1971, and the Philippines in 1977. Given this wide distribution in south-eastern Asia for many years, and the recognition of the species by the CIE identification service in London for many years, there seems little evidence to support the suggestion from Sakimura et al. (1986) that T. palmi may have been confused in the Oriental region with Thrips flavus, another common yellow species, nor the suggestion by Bournier (1983) that some records of Thrips tabaci might have been misidentifications of T. palmi. Despite this, some misidentifications are recorded. Chang (1991) noted that in 1979 T. palmi had been misidentified in Taiwan as an outbreak of T. flavus on cucurbits. Johnson (1986) observed that in Hawaii T. palmi was initially thought to be Thrips nigropilosus until identified by Nakahara et al. (1984). Similarly, T. palmi was misidentified in India as Frankliniella schultzei and under that name considered to be the main vector of tomato spotted wilt disease on groundnut (Palmer et al., 1990). T. palmi has been referred to under at least six other names: Thrips clarus, Thrips gossypicola, Thrips gracilis, Thrips leucadophilus, Thrips nilgiriensis and Chloethrips aureus. Despite this apparent confusion, and subsequent to the redescription of the species by Bhatti (1980), there is currently no essential problem with distinguishing this species from other thrips (Palmer, 1992). As with all small insects, care must be taken when identifying thrips, but recent information and identification systems published on CD-ROM provide detailed colour photomicrographs to facilitate identification (Moritz et al., 2001). Moreover, a similar computer-driven system is available in English, German and Spanish that incorporates a molecular system for identification of larvae and eggs as well as adults (Moritz et al., 2004). Further molecular identification systems can be expected to be developed (Walsh et al., 2005).

AdultsThe structure of the mouthparts of adult females of T. palmi has been examined by Yasumi et al. (1994); the morphology of mouthparts, and the feeding marks on injured leaves indicate that T. palmi is a sap feeder.FemaleColour pale yellow, except antennal segment III usually dark at apex, IV and V usually dark with base pale, VI and VII dark; forewings pale. Antennae with seven segments, terminal segment small; segments III and IV with forked sense cones. Head with no setae directly in front of first ocellus, one pair lateral to first ocellus and a smaller pair nearer the compound eyes; postocellar and postocular setae small. Pronotum with two pairs of long posteroangular setae, remaining setae small; surface with faint transverse lines. Metanotum with median pair of setae not at anterior margin, sculpture transverse at anterior but converging to posterior, paired campaniform sensilla present. Forewing first vein with about 7 setae basally and 2 or 3 setae distally; second vein with a row of about 12 setae. Abdominal tergite VIII posterior margin with a comb of long, fine microtrichia, paired ctenidia present posteromedially from the spiracles; tergite IX with two pairs of campaniform sensilla; tergite II with 4 lateral setae; median tergites with median setae shorter than the distance between their bases, and no sculpture medially, lateral sculpture without microtrichia. Abdominal sternites with 3 pairs of posteromarginal setae, but no discal setae; pleurotergites lacking microtrichia and discal setae.MaleSimilar to female but smaller; tergite II sometimes with only 3 lateral setae; tergite VIII posteromarginal comb often absent laterally; tergite IX setae B1 slightly shorter than, but in line with B2 setae; sternites III - VII each with a large transverse glandular area.LarvaeIn common with other, similar thrips species, T. palmi has two larval stages and two pupal stages. The second-instar larvae can be distinguished from those of other species by details of the sculpture of the dorsal surface, but specimens are even more difficult to prepare for study than are adults. Miyazaki and Kudo (1986) provide an identification key to several species which are common in the Oriental region.

The geographical distribution of T. palmi continues to expand year by year, and the species can be expected to become pantropical in due course. However, establishment of this species is presumably limited by climatic conditions (McDonald et al., 1999). For example, although outdoor overwintering normally occurs in Okinawa (26°N), in the southern part of Kyushu (about 32°N) and further north on mainland Japan, there is no overwintering out of doors, and greenhouses serve as foci for summer populations (Y. Hirose, Institute of Bioloigical Control, Kyushu University, Japan, personal communication). In Australia, infestations are localised, and their extent is probably limited by the prevailing aridity. Strassen (1989) stressed the risk of introduction to Europe.

A record of T. palmi in Wisconsin, USA (Chu et al., 2006) published in previous versions of the Compendium was included in error. Chu et al. (2006) only mentions detection of T. palmi in Georgetown, St. Vincent.

See also CABI/EPPO (1998).

During the second half of the twentieth century T. palmi progressively became a major pest of cucurbits and solanaceous plants, and was introduced and became established in many tropical countries. Walker (1992) lists over 200 records of the plants on which T. palmi has been recorded: aubergine is mentioned in 53 records, cucumber in 29 and watermelon 16. Reports of this thrips infesting Ficus species in the Netherlands resulted in some authors demonstrating that T. palmi could feed on these plants (Loomans et al., 1999) whereas other authors demonstrated that T. palmi did not breed on Ficus (O'Donnell and Parrella, 2005). In 1978 this thrips became a major threat to vegetable growers in Japan, and over the next 10 years spread to an area of about 20,000 hectares (Murai, 2002). By 1990 it had become the most serious pest of cucumber, aubergine and Capsicum annuum both in greenhouses and in open fields in the western part of Japan (Kawai, 1990b). It was detected in glasshouses in the Netherlands in 1988 and 1992 (Parrella and Mound, 1998), and has been intercepted on imported cut flowers from various countries, including Thailand and India.T. palmi is established in South America, and was reported first in Venezuela by Cermeli and Montagne (1993) as attacking Phaseolus vulgaris, potato, aubergine and melon during 1990-91. Other affected crops included cucumber, pepper, sesame, sunflower, soyabean, cowpea, tobacco and squash. In Brazil, Monteiro et al. (1995) recorded T. palmi causing damage to chrysanthemums, aubergines, potatoes, tomatoes and Capsicum. Introduced to Colombia in 1997 it is now a serious pest in that country, and by 2005 it was well established in Guatemala and can be expected to become similarly established in other countries in the area. In Australia, T. palmi was first recorded in Northern Territory on watermelon and in Queensland on cucumber (Houston et al., 1991), but it is not a widespread problem probably due to low humidities. Childers and Beshear (1992) determined (among other thrips species) the occurrence, frequency and distribution of T. palmi in open flowers of citrus in Florida, USA, and in Japan. Miyashita and Soichi (1993) noted that most adults aggregated at the growing point before flower differentiation, and dispersed to other plant parts thereafter. Injury was caused by feeding on leaves around the growing tips before expansion. They found no difference in the portions attacked and in the population density of T. palmi between varieties of chrysanthemum.

Damage by T. palmi is not unlike that caused by many other species of thrips; when populations are high, their feeding causes a silvery or bronzed appearance on the surface of the plant, especially on the midrib and veins of leaves and on the surface of fruit. Leaves and terminal shoots become stunted and fruit is scarred and deformed. Damaged leaves generally show a darkened, glossy, pearly appearance (Bournier, 1987). Johnson (1986) described heavy damage to watermelon foliage as bronzing and total destruction of the vine tips.Bournier (1983, 1987) described damage to cultivated cotton caused by T. palmi and, among the symptoms, observed that the oldest tissue may thicken, warp and finally crackle. Damage to cotton seedlings by T. palmi has also been reported in Thailand by Wangboonkong (1981) if there are long periods of drought early in the season. Apart from Karny (1925), who described T. palmi and observed that it infested both mature and seedling tobacco in Sumatra, one of the earliest reports of damage by T. palmi was Ananthakrishnan (1955). He described the damage to Sesamum plants in Madras, India, as malformation of the stamens, injury to the ovarian wall and the development of a dark pigment on the fruit wall, instead of the usual green colour.Damage has been described by Nakazawa (1981) in Japan as yellowing of the leaves, topping, scratches on the fruits, malformation of the fruits, poor fruiting and death of the whole plant when populations are high. In Martinique, Denoyes et al. (1986) described the damage on the leaves of aubergine, cucumber, melon and other cucurbits. Pantoja et al. (1988) reported severe damage to cucurbits and solanaceous commercial plantings in 1986 in Puerto Rico, where adult and immature thrips fed gregariously on leaves, stems, flowers and developing fruits. Pepper plants became stunted with a bronzed appearance and aubergine plants showed premature fall of developing fruits and buds, and deformed fruits.Kawai (1986b) studied the relationship between the density of T. palmi and the damage to Capsicum annuum and aubergine in Japan. He also studied the relationship between different densities of T. palmi and injury to cucumbers grown in a vinyl house (Kawai, 1986c). The growth of cucumber plants was retarded when thrip numbers were high. The tolerable pest densities were estimated at 5.3 adults per leaf for the total fruit yield and 4.4 adults per leaf for the yield of uninjured fruit (assuming an acceptable yield loss of 5% of the maximum yield). Sakimura et al. (1986) observed that both adults and larvae of T. palmi feed gregariously on leaves, firstly along the midribs and veins. Stems are attacked, particularly at or near the growing tip, and are found amongst the petals and developing ovaries in flowers and on the surface of fruit. They leave numerous scars and deformities, and finally kill the entire plant.

Identification details, illustrated with colour photomicrographs, are given by Moritz et al. (2004), together with a molecular method for identifying larvae and eggs. In the field, T. palmi can easily be confused with several small, yellow species of thrips, such as Thrips flavus, or the pale forms of Thrips tabaci, Frankliniella schultzei, F. occidentalis and T. nigropilosus. T. flavus is generally larger than T. palmi, whereas the yellow species of the common genus Scirtothrips are smaller; moreover the other species mentioned here nearly always have some brown shading on the abdomen, and T. tabaci is unusual in lacking red pigment beneath the ocelli.

As with western flower thrips (Frankliniella occidentalis), the rapid expansion of T. palmi in the last decade of the twentieth century is possibly associated with the development of a new biological strain under the selection pressure of various agrochemicals (Bournier, 1986, 1987), but this suggestion remains to be explored using molecular techniques. DevelopmentAdults usually emerge from pupae in the soil or leaf litter, and move to the young leaves and flowers of the host plant, where they lay their eggs in the green tissue in an incision made with the ovipositor. There are two active larval instars and two relatively inactive 'pupal' instars. At 25°C, the life cycle from egg to egg lasts only 17.5 days (EPPO, 1989). Wang et al. (1989) observed the oviposition behaviour of T. palmi in Taiwan and noted that the pre-oviposition period was 1-3 days for virgin females and 1-5 days for mated ones. Virgin females laid 1.0-7.9 eggs per day, with 3-164 eggs laid during their lifespan. Mated females laid 0.8-7.3 eggs per day and laid 3-204 eggs during their lifespan. The reproductive potential of T. palmi has been studied by Teramoto et al. (1982). In field plantings of cucumber in Hawaii, Rosenheim et al. (1990) noted that the secondary sex ratios were biased towards females. The effect of temperature on population growth was studied by Kawai (1985b) in the laboratory in Japan. At 25°C, the net reproductive rate, female fecundity and daily oviposition rate reached maxima, the values for the last two parameters being 59.6 eggs per female and 3.8 eggs per day, respectively. In Taiwan, the optimum temperature for population growth was found to be 25-30°C, and the number of generations possible in central Taiwan was estimated as 25-26 per year (Huang and Chen, 2004). Likewise Cermeli and Montagne (1993) recorded that at 26°C, on leaves of Phaseolus vulgaris, the life cycle was 11.5 days, the net reproduction rate 18.3, the generation time 27.3 days and the intrinsic rate of natural increase was 0.125 individuals per female per day.Kawai (1988a) analysed the relationship between the density of T. palmi on aubergine or Capsicum annuum and the rate of copulation in plastic houses in Japan. The rate of copulation was low in plots with low densities, and the low-density effect strongly influenced the population fluctuations of T. palmi on both crops. Similar studies were carried out by Kawai (1987) on cucumber.Parthenogenesis (arrhenotoky) in T. palmi has been mentioned by Yoshihara and Kawai (1982) and Bernardo (1991), where the species was confused with T. tabaci. Bernardo (1991) summarized the biology of T. palmi in the Philippines and noted that the highest levels of fecundity (15.6 eggs) and longest adult lifespan (17.4 days) were recorded on watermelon. It was suggested that this could be one of the reasons for the fast build-up in thrips populations and the severe damage on watermelon. Chang (1991) summarized the biology in Taiwan and noted that in southern Taiwan it takes 20-30 days for one generation to develop on cucurbits, with peaks occurring in December and mid-January. Most research activity on the biology of T. palmi is linked to population ecology; there have been few studies on the biochemical aspects of the pest. However, the activity of the enzyme cholinesterase was measured by Kazano and Nishino (1986) in Japan. Rajulu and Gowri (1988) compared the structure and chemical nature of the chorion of eggs of DDT-resistant and susceptible strains of T. palmi from India. They noted that the chorion had three layers; the outer layer was similar in both strains but there were differences between strains in the middle and inner layer of the chorion. Murai (2002) summarized the extensive literature in Japanese concerning T. palmi. Wang and Chu (1986b) described rearing methods in the laboratory of T. palmi in Taiwan. They reared the thrips on pieces of pumpkin leaf in 4.3 x 1.5 cm glass vials. About 61% of the eggs laid developed into adults and development took an average of 13.73 days. Koyama and Matsui (1992) also reared adult T. palmi on dry leaf powder of cucumber, aubergine and tomato.Seasonal PatternsIn Japan, the seasonal prevalence of T. palmi has been investigated by Kajitani et al. (1988) and Sawada (1985) (on aubergine), Matsuno and Okuhara (1985) (on watermelon) and Nishino et al. (1983) (on cucumber and aubergine) in plastic greenhouses. T. palmi is unable to overwinter on outdoor vegetation beyond a northern limit which is mentioned by Sakimura et al. (1986) as 26°N in Japan. Morishita and Azuma (1989) noted that overwintering of T. palmi did not occur in the field at Wakayama in Japan due to low temperatures, but populations were maintained in greenhouses, and by late July adults had dispersed to aubergine fields within a distance of 800 m from the greenhouses where overwintering had occurred. The speed of development of T. palmi at various temperatures in Japan has been investigated by Nonaka et al. (1982). In Taiwan, it was observed by Ho and Chen (1992) that the population of T. palmi peaked in May, and was also suppressed mainly by rainfall. Tsumuki et al. (1987) investigated the survival period of adults and second-instar larvae of T. palmi in the laboratory to determine whether it can overwinter in the field in Japan. It was found that summer populations were susceptible to exposure to low temperatures, while winter populations were more tolerant of them. It is suggested that T. palmi may not be able to overwinter under natural conditions in the area of Okayama. The effect of temperature on population growth was studied by Kawai (1985b) in the laboratory. He established that the threshold temperature for development of T. palmi and the thermal constant for pre-adult stages were estimated to be 11.6°C and 189.1 day-degrees, respectively. The survival rate of larvae and pupae reared on various host plants was evaluated by Kawai (1986d). Cucumber, kidney bean (Phaseolus vulgaris), aubergine and balsam pears (Momordica charantia) had the highest survival rate, followed by melon, pumpkin (Cucurbita maxima) and Capsicum annuum. The generation times were not significantly different between crops, being in the range of 21.2-25.9 days, but there were significant differences in net reproductive rate.Density and PopulationKawai (1983b) explored the relationship between the density of adults on cucumber plants and the number of individuals trapped by sticky traps in a plastic greenhouse in Japan. A positive correlation was found between the adult density per plant and the number of adults trapped. It was concluded that adhesive traps can be used to monitor the relative abundance of T. palmi adults. Kawai (1983b) tested the attractiveness of blue and white for T. palmi. Huang (1989) observed that white was the most effective colour to attract T. palmi to sticky trap plates and that the most suitable height to trap thrips was 0.5 m above the ground. However, Layland et al. (1994) used blue sticky-board traps to monitor T. palmi.A great deal of research has been concentrated on the population growth and distribution of T. palmi on aubergine in glasshouses and in the field in Japan (Kitamura and Kawai, 1983; Kitamura et al., 1984). Morishita and Azuma (1989) noted that in aubergine fields T. palmi was more abundant at field margins than in the centre. Moreover, fences around aubergine fields were effective in preventing immigration of the pest. Nagai (1990a) observed that the population density of T. palmi in fields reached high levels in July and September on untreated aubergines. Kawai (1988b) studied the distribution of T. palmi on aubergine and C. annuum in relation to population density in a plastic greenhouse. He found that on aubergine, adults were most abundant on leaves while larvae were also abundant on flowers. Moreover, more than 99% of the total population infested leaves, while less than 1% infested flowers and fruits, regardless of the population density. Kawai (1983a) noted that adults were found mainly on the middle leaves (5-10 leaves beneath the apical leaf) before topping, and on apical leaves after topping. Larvae were found mainly on the lower leaves (10-15 leaves beneath the apical leaf), on which adults had been present 7 days before. Adults were distributed at random among the plants, but the distribution of the larvae was moderately contagious. On C. annuum, adults were most abundant on flowers, whereas larvae were abundant on fruit. About 50% of adults and 80% of larvae infested leaves, whereas 50% of adults and 20% of larvae infested flowers and fruits, regardless of population density. However, studies by Ho et al. (1993) in Taiwan reported that on aubergine, 52% were on the old leaves, the mid-aged leaves contained 36%, and 12% were on the young leaves. Population densities of T. palmi on aubergine leaves were investigated by Fujioka et al. (1992) before and after planting seedlings in plastic greenhouses in Japan. A positive correlation was found between the densities before and after planting, indicating that effective control before planting would suppress the pest after planting. Mulching ridges with polyvinyl film improved the effects of spraying insecticides. Populations of T. palmi on C. annuum in vinyl houses in Japan were investigated by Morishita and Azuma (1988) during 1985-86. Seasonal changes were similar in most of the houses, densities being low in winter, increasing in spring, and peaking in April or May.In Hawaii, Rosenheim et al. (1990) observed that the density of T. palmi was greatest on the foliage of cucumber and lowest on the fruits. In Taiwan, Huang (1989) observed that annual populations of T. palmi on wax gourd peaked in late April and May and again in early October and November. Su et al. (1985) studied adult populations of T. palmi in two areas in Taiwan on aubergine. Populations peaked in April and July in one area and April, July and October in the other. They also found that population densities were influenced by humidity, temperature, rainfall and the duration of sunshine.Rainfall is known to have an effect on thrips populations; Etienne et al. (1990) noted that rainfall depressed populations of T. palmi in Guadeloupe. They also noted a drop in density of T. palmi in December and attributed this to increased precipitation and the termination of chemical treatment in the middle of November. In Trinidad, Cooper (1991b) also observed that infestation levels were significantly reduced by rain. Intercropping a potato crop with Allium cepa or A. sativum in West Java increased populations of T. palmi if spacing was less than 0.75 m (Potts and Gunadi, 1991).The movement of adult T. palmi on cucumber in a greenhouse in Japan was investigated by Kawai (1986a). He compared the differences between cucumbers that were covered with vinyl film that absorbed light in the ultraviolet region of the spectrum as opposed to a standard vinyl film. Dispersal and flight in the plot covered with the absorbent film were less than for the standard film. Immigration into the plot covered with absorbent film was less but emigration more for the standard film. Kitamura and Kawai (1984) studied the distribution pattern of T. palmi on aubergine in a greenhouse in Japan. In India, panicles of mango were sampled by Verghese et al. (1988) for T. palmi; their distribution was found to be aggregated, and the data fitted a negative binomial model.CompetitionInterspecific competition between T. palmi and Aphis gossypii on aubergine was studied in the laboratory and in the field in Japan (Kawai, 1985a). The population of T. palmi became extinct regardless of the initial densities of the two species. Moreover, when the aphid was released onto aubergines in the field, the population density of T. palmi also decreased rapidly.Plant AssociationsThe removal of weeds had a beneficial effect on T. palmi populations in Trinidad (Cooper, 1991b), suggesting that weeds might be alternative hosts, but Nagai and Tsumuki (1990) tested T. palmi-infested weeds in an unheated greenhouse in search for winter host plants of the pest. The weeds included Vicia sativa, Cerastium glomeratum and Capsella bursa-pastoris. They discovered that at least one generation was produced on V. sativa and C. glomeratum, but no reduction of the adult population density on V. sativa was observed under temperatures as low as -3 to 7°C from mid-January to mid-February.T. palmi as a VectorThe specialized mouthparts are adapted for sucking (Lewis, 1973). T. palmi (initially misidentified as Frankliniella schultzei) is the main vector of tomato spotted wilt disease on groundnut in India (Palmer et al., 1990). The current status of research on the groundnut disease caused by Peanut bud necrosis virus, and its transmission by T. palmi, is reported by Buiel et al. (1993, 1995). Recent advances in the determination of genome structure, host range, transmission and spread of tospoviruses, the epidemiology of peanut bud necrosis and resistance to both the vector and the virus are reported. T. palmi also transmits Tomato spotted wilt virus on watermelon in Japan (Kameya-Iwaki et al., 1988; Honda et al., 1989) and Taiwan (Yeh et al., 1992; Chen et al., 2004), and is a vector of Calla lily chlorotic spot virus in Taiwan (Chen et al., 2005).

Walker (1992, 1994) has reviewed the pest status of T. palmi; much of the information given here is from these reviews, together with later information. The importance of the pest on vegetable crops in South-East Asia was emphasized by a workshop held in Bangkok, Thailand (Talekar, 1991) where seven of the eight papers presented listed T. palmi as causing concern for vegetable growers in their region. However, research on this species attacking vegetables in Taiwan was reduced when it was realised that it is effectively a 'resurgence pest', that is, populations tend to increase disproportionately following heavy use of insecticides. The reasons for this phenonmenon require further study. The economic injury density in Japan has been estimated at 0.105 adults per flower or 4.4 adults per sticky trap per day on Capsicum annuum in vinyl houses (Morishita and Azuma, 1988), assuming an acceptable yield loss of 5% of the maximum yield. Kawai (1986b) also reported that economic injury thresholds were low in vinyl houses in Japan, assuming an acceptable yield loss of 5% of the maximum yield, with 0.08 and 4.4 adults per leaf for aubergine and cucumber, respectively, and 0.11 adults per flower for C. annuum. Morishita and Azuma (1989) considered counting injured fruits to be a better sampling method than counting insects on leaves.In the Philippines, plantings of aubergine intended for seed production had to be abandoned due to severe T. palmi damage and even the application of insecticide as often as every 4 days failed to provide satisfactory control (Bernardo, 1991). Chang (1991) lists T. palmi as one of Taiwan's most important pest thrips; damage was first observed on cucurbits in 1979, but the species was incorrectly identified as T. flavus. T. palmi has also been identified as a important pest of potato in Taiwan by SEAMEO SEARCA (1991).However, Bournier (1986) reported that T. palmi caused insignificant damage on cotton, tobacco and wild plants in Java, Sumatra and India. Miyazaki et al. (1984) also observed, during a survey of soyabean in Java, that T. palmi did not cause heavy damage except in one instance on aubergine. Cooper (1991b) recorded infestations of 300-700 T. palmi per leaf on aubergine and cucumber, resulting in crop losses of 50-90% in Trinidad. He suggested that T. palmi may have been brought to Trinidad in the winds of a tropical depression during 1988, but it has also been postulated that it may have gained entry through plant material from another Caribbean island, for example Martinique, where it is reported as a serious pest. Pantoja et al. (1988) noted that the climatic conditions in Puerto Rico are favourable for the early development of large populations of T. palmi on commercial crops as well as on weeds. Guyot (1988) reported the disastrous economic effect that T. palmi had in Guadeloupe when aubergine exports fell from 5000 tonnes in 1985 to 1600 tonnes in 1986, and in Martinique where 37% of the vegetable crops of the two main co-operatives were attacked by T. palmi, including 90% of aubergine crops.In Hawaii, Johnson et al. (1989) observed that, together with Aphis gossypii, T. palmi was the major foliar pest on Oahu (1984-85). Welter et al. (1989) studied mixed infestations of T. palmi and the western flower thrips, Frankliniella occidentalis, and noted significant reductions in total cucumber yield, mean fruit size and total fruit. The population trends of T. palmi on commercial watermelon plantings in Hawaii were surveyed by Johnson (1986). Peak infestation levels varied from 2.5 to 53.6 individuals per leaf and from 18 to 97% infested vine tips per planting. Johnson (1986) pointed out that T. palmi could establish itself in the continental USA, given the extensive flow of air traffic between Hawaii and the mainland, especially California, but it was not until 1991 that T. palmi was found in the USA, not in California as predicted by Johnson but in Florida (FAO, 1991). Heavy infestations were detected on potato, aubergine, Capsicum, Phaseolus vulgaris, yellow squash and several weeds. The likely economic impact of this pest if it became established in greenhouses in the UK was considered to be very severe, with a benefit to cost ratio for one eradication campaign being as high as 110:1 (MacLeod et al., 2004).

T. palmi is not easily detectable because of its small size, so quarantine procedures are difficult to manage and this pest has probably slipped through the net with increased traffic in plant produce around the world.Sakimura et al. (1986) observed that both adults and larvae of T. palmi feed gregariously on leaves, firstly along the midribs and veins. Stems are attacked, particularly at or near the growing tip, and they are found amongst the petals and developing ovaries in flowers and on the surface of fruit. They leave numerous scars and deformities, and finally kill the entire plant. Ho et al. (1993) reported that old aubergine leaves are a good site for sampling in Taiwan.Layland et al. (1994) described methods used to monitor T. palmi, using blue sticky-board traps and water-tray traps, in Northern Territory, Australia.Kawai (1983b) explored the relationship between the density of adults on cucumber plants and the number of individuals trapped by sticky traps in a plastic greenhouse in Japan. A positive correlation was found between the adult density per plant and the number of adults trapped. He concluded that adhesive traps can be used to monitor the relative abundance of T. palmi adults. Kawai (1983b) tested the attractiveness of blue and white for T. palmi. Huang (1989) observed that white was the most effective colour to attract T. palmi to sticky trap plates and that 0.5 m above ground level was the most suitable height to trap thrips. Layland et al. (1994) used blue sticky-board traps to monitor T. palmi.

Chemical Control

Chemical control has given a measure of protection in many countries, for example, on cucurbits in New Caledonia (Guterierrez, 1981) and on watermelon in Hawaii (Johnson, 1986). Nine insecticides were tested against T. palmi on potato in Mauritius, but none gave satisfactory results (Anon, 1987b). Seal et al. (1994) tested the effectiveness of five insecticides for the control of T. palmi on Phaseolus vulgaris, Cucurbita pepo and aubergine in Florida during 1993. Some insecticide screening has also been recorded from South America. In Brazil, Cermeli et al. (1993) tested 11 insecticides and observed a high level of tolerance to chemical control. Flufenoxuron, imidacloprid and chlorfluazuron were the most effective insecticides; however, no insecticide was more than 81.5% effective.

Nishino (1987) studied seasonal fluctuations of susceptibility to insecticides in Japan. Nagai et al. (1981) investigated the effect of paired insecticide mixtures, and Nishino et al. (1982) reported on the effects of certain insecticides on the control of T. palmi. Yoshihara et al. (1984) provided a method of evaluating the susceptibility of T. palmi to several insecticides. Morishita and Azuma (1989) noted that population densities changed in response to pesticide applications in fields of aubergine. Nemoto (1995) evaluated the side effects after aubergine plants were continuously sprayed with the insecticides permethrin, milbemectin, phenthoate and imidacloprid in Japan. The effects on populations of pests and their natural enemies were assessed and revealed the importance of natural enemies such as Orius species. Imidacloprid, which was highly effective against Hemiptera and T. palmi, caused a resurgence of Tetranychus kanzawai. Milbemectin, which had a minimal adverse effect on Orius spp., when used in combination with imidacloprid, maximized the latter's advantages while minimizing its disadvantages. A new insecticidal and acaricidal compound, PF1018, has been reported by Gomi et al. (1994). Application by leaf dipping caused 96% mortality to second-instar larvae of T. palmi.

Nakazawa (1981) reported that T. palmi was not susceptible to acephate, phenthoate or fenitrothion and that no insecticides were practical for use against T. palmi, but by 1992 fenobucarb was being used.

In Taiwan, Su et al. (1985) found that deltamethrin and cypermethrin were effective in controlling T. palmi on aubergine, and Huang (1989) reported that pesticides were most effective if sprayed in the morning or evening. In Hong Kong, Riddell-Swan (1988) highlighted the fact that T. palmi had become resistant to almost all organophosphates.

Etienne et al. (1990) discovered that populations of T. palmi on aubergine in Guadeloupe were much higher in diazinon- and profenofos-treated plots than untreated plots; this was attributed to reduced predator activity in the treated plots. Denoyes et al. (1986) commented that in Martinique, where T. palmi had been discovered in 1985, no chemical control measures had so far been successful, but by 1989, Bon and Rhino reported that profenofos and abamectin were effective.

On Guam, Schreiner (1991) noted that there was no significant effect on the yield of cucumbers in the field, with up to 80 T. palmi per leaf after treatments with carbaryl, dimethoate and Bacillus thuringiensis.

Insect Growth Regulators

The effects of the juvenile hormone mimic pyriproxyfen on T. palmi were evaluated by Nagai (1990c); no difference in mortality was found between those reared on treated or untreated leaves. However, during the pupal stage mortality rose to 70% compared with 30% on untreated leaves. There was no difference in the hatchability of Orius sp. in the laboratory when treated with pyriproxyfen. In the greenhouse, populations of Orius sp. were reduced to very low levels within 10 days of application of carbaryl alone and in combination with pyriproxyfen, but its population density was the same as that on untreated plants when pyriproxyfen was applied on its own.

Nagai et al. (1988b) showed that flufenoxuron inhibited the ecdysis of first-instar nymphs and metamorphosis of second-instar nymphs to pupae, but did not affect the survival rate and fecundity of females. The effect of flufenoxuron on populations on aubergine in the field in Japan was examined in September and October, when suppression of the population by flufenoxuron was slower than by sulprofos.

The activity of five chitin-synthesis inhibitors against T. palmi infesting melon in Japan was evaluated in the laboratory and greenhouse. Diflubenzuron, teflubenzuron, chlorfluazuron, flufenoxuron and chiromazine were tested in the laboratory. The number of larvae which dropped from leaves in order to pupate decreased significantly, and only larvae treated with chiromazine pupated. In the greenhouse, chlorfluzuron and flufenoxuron controlled T. palmi more effectively than diflubenzuron and conventional sprays of fenobucarb (Kubota, 1989).

Integrated Pest Management

Hirose (1991) stressed the importance of considering biological control for T. palmi and also suggested that the resurgence of this pest in South-East Asia during the last 10 years was due to the elimination of its natural enemies by repeated insecticide applications. The effect of insecticides on Orius sp. was investigated by Nagai (1990b) in Japan. Eggs were treated by dipping and results showed that chinomethionate, bromopropylate, pirimicarb and phosalone showed low toxicity. However, carbaryl, a mixture of malathion and fenobucarb, phenthoate and fenthion showed high toxicity to the eggs. Orius apparently showed no susceptibility to buprofezin or bromopropylate when aubergines were sprayed in the field for control of T. palmi, but phosalone, chlorfluazuron and flufenoxuron were highly toxic to Orius species. Nagai (1990a) used fenthion as a control measure when evaluating the effects of predation by Orius. Further evaluation was presented by Nagai (1991a, 1992).

Kawai and Kitamura (1987) concluded that control of T. palmi using insecticides alone was difficult because of the need for the safe use of agricultural chemicals, and they suggested that an integrated control system should be established. A population model of T. palmi was constructed to evaluate the effectiveness of various control methods and to develop an effective control system for T. palmi on cucumber cultivated in a plastic greenhouse.

Kawai and Kitamura (1987) evaluated various control measures on cucumber in plastic greenhouses in Japan and recommended an integrated pest management system. Integrated control of T. palmi on cucurbits in the field in Japan is discussed by Suzuki et al. (1986). An integrated pest management system has already been implemented in Martinique and Guadeloupe (Denoyes et al., 1986).

A pest management strategy was suggested by Hata et al. (1991) in Hawaii to separate susceptible cultivars from preferred cultivars of Dendrobium and the wild bamboo orchid Arundina graminifolia on which T. palmi feeds. Cooper (1991a) provides recommendations for the integrated control of T. palmi in vegetables in Trinidad and Tobago. IPM is probably the most important strategy for dealing with T. palmi, particularly in view of the recognition in Taiwan that frequent insecticidal treatments can lead to increased populations.

Post-Harvest Treatment

There have been few reported efforts directed towards post-harvest treatments to eliminate T. palmi from plant material. Jacqua and Etienne (1987) dipped aubergine fruit in water at various temperatures after harvest to eliminate T. palmi under the calyx. Dipping at 45°C was better for subsequent fruit conservation, as treatment at 50°C could induce fruit damage.

Mann et al. (1995) evaluated post-harvest treatments to control T. palmi on Dendrobium orchid blossoms in Hawaii, including insecticidal dips, isopropyl alcohol dips, insecticidal fogs and hot-water immersion. The limiting factor for all post-harvest treatments was phytotoxicity, characterized by a loss of vase life that differed among cultivars. Insecticidal dips and insecticidal fogs were less phytotoxic than hot-water immersion and isopropyl alcohol dips. Hata et al. (1993) also experimented with insecticidal dips in Hawaii. Double dips applied after harvest reduced >95% of a mixed infestation of Frankliniella occidentalis (90%) and T. palmi (10%) infesting Dendrobium 'Uniwai Princess' blossoms, compared with untreated controls. Single insecticidal dips were not as effective.

Feeding Deterrents

Although Yasuda and Momonoki (1988) noted that a variety of aubergine resistant to T. palmi has been introduced from South-East Asia into Japan, the majority of research on feeding deterrents has been conducted on tomatoes. The effect of tomato leaf constituents on the survival of T. palmi was described by Yasumi et al. (1991). Adult females survived for a long period on a filter paper disc soaked in a 3% aqueous sucrose solution, even if the sucrose-impregnated disc had methanol extracts of cucumber or aubergine leaves applied to it. However, when methanol extract of tomato leaves was applied to the sucrose-impregnated disc, all of the the insects died within a few days. It was concluded that the active substance was an antifeedant rather than a toxin.

Athough T. palmi is polyphagous, Hirano et al. (1994) stated that it does not attack tomatoes in Japan. A crystalline compound possessing strong antifeedant activity was isolated from tomato leaves and was identified as the steroidal glycoalkaloid alpha-tomatine. It was concluded that the immunity of tomato plants to T. palmi is explained solely by the occurrence of alpha-tomatine, because T. palmi does not use airborne information (attractants/repellents) to avoid tomato leaves, and tomato leaves apparently provide the required phagostimulants and nutrients.

Short-distance walking responses of T. palmi to non-preferred tomato and preferred aubergine leaves have been examined by Hirano et al. (1993). In leaf-disc assays, adult females of T. palmi moved randomly and there was neither an oriented approach towards aubergine leaves nor an active avoidance of tomato leaves. Similar results were obtained on leaf-extract-treated filter paper discs. It was concluded that T. palmi is not deterred from tomato during the period from searching to initial settlement; this is supported by a report by Cermeli and Montagne (1993) who observed that, although T. palmi was collected on tomato crops, it did not cause economic damage.

Biological control

Predators
Anthocorid bugs play the most significant role in the natural control of T. palmi in many areas where it is a pest. Seven species have been mentioned in the literature: Orius sp. in Japan; O. similis and O. tantillus in the Philippines; O. sauteri in Taiwan; O. maxidentex and Carayonocoris indicus in India; O. insidiosus in Hawaii, USA; and Bilia sp., O. minutus, Wollastoniella parvicuneis and W. rotunda in Thailand. Most of these investigations have been conducted in Japan and, in most cases, the species used has not been identified and is referred to only as Orius sp. Until the identification of the species is settled, it is difficult to make comparisons or suggest other species which might be considered in biological control projects.

The biology of Orius sp. from Japan has been studied by Nagai (1989), who determined the duration of the egg and larval stages in the laboratory. Using T. palmi as prey, the author found that the duration of the stages decreased with an increase in temperature. The predatory effect of Orius sp. on the density of T. palmi was investigated on potted aubergine in a screenhouse (Nagai et al., 1988a) and on aubergine in the field (Kawamoto and Kawai, 1988; Nagai, 1990a). It was concluded that the introduction of Orius sp. lowered the population density of T. palmi on aubergine. Conversely, the population densities of T. palmi, Tetranychus kanzawai and Tetranychus urticae were greater when populations of Orius had been eliminated by insecticides (Kawai and Kawamoto, 1994). Orius spp. were active from May until November, with two population peaks in July-August and in September.

The dispersal of Orius spp. was evaluated by Kawai (1995) on greenhouse-grown aubergine infested with T. palmi. The population density of T. palmi decreased on nine plants adjacent to the plant where the predator was released within a few days after release, and remained low until the end of the examination. Although individuals of Orius spp. dispersed to other parts of the greenhouse afterwards, they could not control the population density of T. palmi effectively, due to the delay in their arrival. It took about 1 month for Orius spp. individuals to disperse to the other end of the greenhouse. It was concluded that the dispersal ability of early-instar nymphs is low while that of late-instar nymphs and adults is high.

Kawamoto and Kawai (1988) observed the effect of Orius sp. on populations of arthropod pests on aubergine in the field in Japan in 1987 and noted that populations of Orius sp. were lower in insecticide-treated plots than in untreated plots. They also noted that populations of T. palmi (together with Tetranychus kanzawai) were higher in treated plots, and suggested that Orius sp. was effective in reducing populations of arthropod pests on aubergine.

In China, Wei et al. (1984) studied the biology and predatory behaviour of O. similis. They found that in the laboratory, one individual of O. similis could prey on ca 440 individuals of T. palmi during its lifetime (both as a nymph and adult). Nagai (1991b) also studied Orius under laboratory conditions in Japan and noted that at 25°C, one average adult female of Orius sp. preyed on 22 second-instar larvae or 26 adults of T. palmi in 24 hours. However, it was further noted that Orius adults did not eat the eggs of T. palmi.

O. tantillus was studied in the laboratory in the Philippines by Mituda and Calilung (1989). The duration of the egg stage averaged 4.52 days and the total life cycle was 14.76 days for males and 16.52 for females. Adult Orius sp. consumed up to 20 thrips per day and the total number killed throughout the predators' lifetime averaged 205.71 for males and 228.10 for females. It was noted that laboratory studies have demonstrated the great potential of the anthocorid as a biological control agent against T. palmi. Six weed species and five crop plants have been recorded as host plants for O. tantillus in the field. Bernardo (1991) lists numerous predators that have been found associated with T. palmi in the Philippines.

The predatory mites Amblyseius mckenziei and A. okinawanus, and Orius sp., were investigated by Kajita (1986) in Japan. Their prey stage preference, prey consumption and feeding behaviour was studied in detail. Adult females of the two species of mites preferred first-instar larvae as prey to second-instar larvae and adult thrips, whereas the numbers of first- and second-instar larvae consumed by second-instar Orius sp. did not differ greatly. Orius sp. did not differ significantly in the number of prey consumed from the two species of mites.

Kumar and Ananthakrishnan (1984) studied the anthocorids O. maxidentex and Carayonocoris indicus in the laboratory and in the field near Madras, India. O. maxidentex fed on T. palmi on the young foliage of sesame and, after the crop was harvested, was abundant on the weed Croton sparsiflorus, preying on T. palmi until prey populations died out in September.

In Trinidad the only predator discovered was the coccinellid beetle Coleomegilla maculata (Cooper, 1991b). In Hawaii, Mau et al. (1989) recorded O. insidiosus and Franklinothrips vespiformis as predators of T. palmi, and in Taiwan, Wang (1994) evaluated the mirid bug Campylomma chinensis and O. sauteri in aubergine fields. Population densities of the mirid were higher than those of the anthocorid.

Hirose (1990) explored the possibilities of using natural enemies against T. palmi in South-East Asia and Japan. The anthocorid predator Bilia sp. from Thailand was suggested to control T. palmi in Japan. The periodic release and conservation of predators was also discussed. There is increasing interest in anthocorids attacking T. palmi in Thailand, and Yasunaga and Miyamoto (1993) reported three species associated with T. palmi in aubergine gardens in Thailand; Wollastoniella rotunda is described as new and O. minutus and O. tantillus are recorded from this region for the first time. A new subgenus, Paraorius, is proposed for Orius tantillus. Two species, Wollastoniella parvicuneis and W. rotunda, were noted by Yasunaga (1995) as preying on T. palmi in northern Thailand.

Parasitoids
There have been few attempts to use hymenopterous parasitoids for biological control of T. palmi. Hirose (1990) suggested the introduction of an unidentified eulophid larval parasitoid, Ceranisus sp., to control T. palmi in Japan. Hirose (1991) offered a convincing argument to explore further the biological control of T. palmi using Ceranisus menes, a eulophid parasitoid native to Japan and Hirose et al. (1993) reported C. menes on aubergine in home and 'truck gardens' in Japan and commented on the biology of the parasitoid and rates of parasitism.

Fungal pathogens
Saito et al. (1989) recorded the entomopathogenic fungus Neozygites parvispora for the first time on T. palmi on melon in a greenhouse in Japan. About 10% of adults and nymphs were infected, but the fungus did not control the pest population. Studies were carried out on the fungus Beauveria bassiana, isolated from T. palmi in Japan; its pathogenicity and the effect of pesticides (including fenobucarb) on hyphal growth was observed.

Hall (1992) reported a new pathogen of T. palmi in Trinidad. Approximately 80% of a population of T. palmi on an abandoned aubergine crop in Trinidad and Tobago was found to be infected by a fungus of the genus Hirsutella. This appeared to be the first deuteromycete pathogen found on T. palmi and isolated in pure culture.

The control of T. palmi by a mycoinsecticidal preparation of Verticillium lecanii was reported by Saito (1992). When a preparation of V. lecanii was applied four times weekly to melons in glasshouses in Japan, population densities of the pests were maintained at low values, compared with rapid increases observed in an untreated glasshouse.

Host-Plant Resistance

Matsui et al. (1995) screened plants from the genus Solanum for resistance to T. palmi. Four wild species, S. viarum, S. sisymbrifolium, S. nigrum and 'IK-35' (an unknown species) were identified as resistant.

Cultural Control

In the Philippines, Ruhendi and Litsinger (1979, 1982) and Litsinger and Ruhendi (1984) studied the effect of rice stubble and straw mulch on the suppression of preflowering insect pests (including T. palmi) on cowpeas (Vigna unguiculata) sown after puddled rice. They hypothesized that rice stubble and straw mulch, by covering bare soil, interfere with visual cues used by migratory thrips and leafhoppers to locate a favourable habitat.

Riddell-Swan (1988) found that mulching beds of hairy gourds with white cloth and black polythene reduced thrips populations, and surrounding the field with reflective material not only reduced the pest infestation but also improved yields. Suzuki and Miyara (1984) also studied agricultural covering materials in Japan. Nasu et al. (1986) experimented with the effect of covering aubergine with transparent plastic film in a plastic greenhouse in Japan. They found that density was low compared to the control, and the effect lasted long after the removal of the film. The invasion of thrips was effectively prevented by using cheese cloth, but the quality of fruit was low because of the raised air temperature.