T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 1 Chapter General introduction _________________________________________________________________________ 1.1 The western flower thrips Thrips are small, slender, usually winged, insects belonging to the order Thysanoptera, which is composed of two sub-orders, the Terebrantia and the Tubulifera. The seven families of the Terebrantia, the Uzelothripidae, Merothripidae, Aeolothripidae, Adiheterothripidae, Fauriellidae, Heterothripidae and Thripidae, account for approximately 2065 of the 5000 or so species in this order (Mound, 1997). The remaining 3100 species belong to the single Tubuliferan family Phlaeothripidae. The majority of species are phytophagous and some of these are serious pests of protected crops, whilst a few species are mycophagous or predatory in their feeding habits. Pest species occur in the four subfamilies of the Thripidae (the Thripinae, Panchaetothripinae, Dendrothripinae and Sericothripinae) and in phlaeothripid subfamily Phlaeothripinae. Since the mid 1980’s the western flower thrips, Frankliniella occidentalis (Pergande), a highly polyphagous species, has spread from its native western USA to a worldwide distribution. This species is currently a major pest of both horticultural and floricultural protected crops in the UK. Cucumbers and chrysanthemums are particularly affected. 1 T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 1.2 Biology In common with the majority of Terebrantia, the F. occidentalis life cycle has six developmental stages: the egg, two feeding larval stages, two non-feeding pupal stages and the adult stage (Bryan & Smith, 1956; Lewis, 1973). The development of thrips is neither truly holometabolic or hemimetabolic. Although the appearance of larval stages closely resembles the adults, suggesting hemimetabolic development, the pupal stages undergo major internal reorganisation, which is typical of holometabolic development. Whilst thrips development is intermediate between holometabolism and hemimetabolism, the majority of researchers refer to thrips as holometabolic insects (Lewis, 1973). 1.2.1 Egg The F. occidentalis are haplodiploid and reproduce by arrhenotokous parthenogenesis (Bryan & Smith, 1956; van Rijn et al., 1995). Unfertilised haploid eggs will result in male offspring, and fertilised diploid eggs in female offspring, although mated females may produce both male and female offspring. The kidney-shaped egg has a smooth exterior, is pale white in colour, approximately 200 μm in length, and is laid directly into plant tissue (Moritz, 1997). During embryogenesis, the cells that would form the right mandible die and, when the egg is about to hatch, red eyespots can be clearly seen through the egg chorion. At the anterior end of terebrantian eggs there is an operculum, which is removed by a saw-shaped oviruptor at hatching. The larva emerges in an embryonic cuticle. In Kakothrips pisivorus (Westwood) this embryonic cuticle splits by movements of the pronotum and is then forced down to the tip of the abdomen by peristaltic movements, exposing the antennae and legs (Kirk, 1985). The larva, which is still attached to the egg by the tip of the abdomen, then rotates its body until the legs come into contact with the substrate, enabling the larva to detach itself from the egg (Lewis, 1973). This developmental stage requires approximately 2.6 d at 25°C (van Rijn et al, 1995). 1.2.2 Larvae There are two larval stages in the F. occidentalis: larvae I and larvae II, which require approximately 2.3-2.8 d and 3.6-3.8 d to complete at 25°C, respectively (van Rijn et al, 1995). After hatching, larvae I are translucent, but soon develop a yellow colouration. Late larvae I and early larvae II are very similar in appearance, and can only be distinguished microscopically. Larvae I possess one pair of setae on the sternites of segments IV-VIII, whilst larvae II have three pairs of setae on these segments. Similarly, 2 T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 the sex of larvae can only be distinguished microscopically. Female larvae I have three pairs of setae on abdominal segment IX, whilst males have four pairs. In larvae II, females have five pairs of setae and males have six pairs (Heming, 1991). Larvae are able to walk and feed immediately after hatching. Walking is aided by arolia, or foot bladders, on each leg at the end of the tarsi. The arolia can be extruded and retracted, to grip the substrate, enabling larvae to walk on very smooth surfaces. The mouthparts are grouped into a structure called the mouthcone, which protrudes from beneath the head. Thrips use the ‘punch and suck’ mode of feeding. The left mandible is used to pierce a plant cell, creating a hole through which the maxillae, which form a tube, are inserted to suck up the cell contents (Chisholm & Lewis, 1984). Although larvae, being soft bodied, appear vulnerable to attack by predators, they are capable of mounting a defence. F. occidentalis larvae repel attacks by vigorously jerking and wagging their abdomen at the predator. This defence is probably most effective in larvae II (van der Hoeven & van Rijn, 1990), as is the case in Thrips tabaci Lindeman (Bakker & Sabelis, 1989). In addition to defensive abdominal movements, the larvae of many thrips species will usually produce a drop of clear fluid from the tip of the abdomen when attacked. In the F. occidentalis, this anal droplet (AD) contains decyl acetate and dodecyl acetate, which are reported to function as an alarm pheromone, causing nearby larvae to move away from the emitter of the AD (Teerling, 1992; Teerling et al., 1993). 1.2.3 Pupae Late larvae II of the F. occidentalis leave the host plant and pupate in the soil underneath (Bennison et al., 2001) at a depth of 1-5 mm (Helyer et al., 1995). There are two quiescent pupal stages in the F. occidentalis: propupae and pupae, which require 1.1-1.2 d and 2.6- 2.8 d to complete at 25°C, respectively (van Rijn et al, 1995). Propupae are distinctly different from larvae II. They are paler, possess wing buds, have erect antennae and are almost unable to walk, although they will move a little when disturbed. Pupae are distinct from propupae. The antennae are laid flat against the head and the wing buds, which are longer than in propupae, are initially clear but darken shortly before emergence of the adult. Unlike larvae, the pupal stages do not produce an anal droplet in response to attack, although they jerk the abdomen in a manner similar to that in larvae (personal observation). 3 T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 1.2.4 Adults Upon hatching, both adult males and females have a pale colouration (Bryan & Smith, 1956). Females are inactive for the first 24 h, and within 48 h have undergone teneral development which is probably temperature dependent (Lewis, 1973), and gained their final colouration. Whilst males remain pale in colour, females will develop one of three different colour forms (Bryan & Smith, 1956). According to Bryan & Smith (1956), these colour forms are under genetic control, the pale form being dominant, the dark form recessive and the intermediate form heterozygous. Sexual dimorphism is apparent, with females being larger than males. After completion of teneral development, the adults of F. occidentalis are extremely active, both in walking and in flight. The flight periodicity, in Portuguese glasshouses, for males and females is the same with one peak of flight activity mid-morning and another mid- afternoon (Mateus et al., 1996). There are no reports on the genetics of this periodicity in the literature. Higgins (1992) found that, at low population densities, 70-90% of thrips caught in traps are male, but at high densities this is reversed, with 65-90% of trapped thrips being female. Very few male F. occidentalis are found on the flowers and leaves of crops, 2-12% and 1-3%, respectively, even when high numbers are being trapped. The reason for this is unclear. It may be that males only remain on flowers long enough to mate with females, which concentrate in cucumber flowers (Higgins & Myers, 1992) and are more likely to be trapped as they spend more time flying (Higgins, 1992). Females do not begin to lay eggs until approximately 1.8 d after eclosion at 25°C (van Rijn et al., 1995). Oviposition rate peaks within the first 10 d after eclosion and then declines gradually, although the number of eggs laid is influenced greatly by diet (Trichilo & Leigh, 1988). Oviposition is known to occur mainly during daylight (Kirk et al., 1999; Kiers et al., 2000) and, in cucumber crops, most eggs are laid in the leaves, with younger leaves being selected over older leaves (de Kogel et al., 1997). Few eggs are laid in the plant stems or flowers, which are ephemeral (Kiers et al., 2000). The longevity of females thrips is longer than that for males (Lewis, 1973). In F. occidentalis this can be as long as three months at 15°C, although this reduces with increasing temperature (Katayama, 1997). 4 T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 1.3 Economic impact 1.3.1 Introduction to the UK The F. occidentalis began to spread within the USA in about 1980 and its presence was detected in Europe, in The Netherlands, in 1983 (Mantel & van de Vrie, 1988; Baker et al., 1993). In June 1986, an outbreak of F. occidentalis was confirmed on chrysanthemums in Cambridge at a single location. Phytosanitary measures were initiated by the Ministry of Agriculture, Fisheries and Food (MAFF) in an attempt to eradicate the infestation, but this strategy soon changed to one of limiting the spread. In March 1987 the first infestation of a cucumber crop was confirmed and by the end of that year, 154 horticultural sites were confirmed to have F. occidentalis. In June 1989, MAFF decided that the F. occidentalis should be classed as established in the UK, and ceased attempts to limit its spread (Baker et al., 1993). 1.3.2 Crops affected in the UK In California, F. occidentalis have been collected from plants belonging to almost every order of Spermatophyta (Bryan & Smith, 1956). In the first two years or so of the presence of F. occidentalis in the UK, this species had been found on 69 genera of plants (Baker et al., 1993). Although F. occidentalis may be able to survive mild winters outside glasshouses in the UK (McDonald et al., 1997), they are primarily a pest of crops grown in glasshouses, and are considered to be of major importance in cucumbers (figure 1.1), sweet peppers, chrysanthemums and bedding plants (Jacobson, 1997). Control of F. occidentalis is largely through Integrated Pest Management (IPM, §1.3.4), and this is estimated to cost cucumber and sweet pepper growers £15,000 ha-1 season-1 and £7,200 ha-1 season-1, respectively (Jacobson, 1997). Of these costs, between approximately 20-30% are accounted for by biological control agents against thrips (§1.3.4.1). 1.3.3 Damage caused by F. occidentalis F. occidentalis can affect crops, such as cucumber and sweet pepper, either directly, through feeding and oviposition activity, or indirectly, through the transmission of plant viruses. Feeding and oviposition activity can affect the crop through activity on the fruit, and through activity on the rest of the plant that ultimately has a detrimental impact on the crop. Adult male and female F. occidentalis exhibit differential feeding behaviour (van de 5 T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 Wetering et al., 1998). Females feed more frequently and intensively than males, and so cause more scar formation on fruit. In cucumbers, this can result in curvature of rapidly growing fruit (figure 1.2), and bronzing and silvering in both cucumbers (figure 1.3) and sweet peppers, resulting in significant downgrading of fruit and so representing a financial loss to the grower (Shipp et al., 1998; Hao et al., 2002). F. occidentalis activity on the rest of the plant can reduce plant growth, photosynthesis (figure 1.4) and marketable yield, although such effects are not as serious as the effects on fruit (Shipp et al., 1998; Hao et al., 2002). In both cucumber and sweet pepper crops, plants are slow to recover from thrips damage once yield loss has occurred, highlighting the importance of monitoring pest populations (§1.3.4.3). Indirectly, F. occidentalis are vectors of two commercially important plant viruses: tomato spotted wilt virus (TSWV) and impatiens necrotic spot virus (INSV). The transmission of TSWV and INSV by F. occidentalis can have a greater impact than the direct effects that this species may have on a crop. Whilst F. occidentalis can be controlled through IPM and serious crop damage avoided, plants infected with these viruses cannot be treated. Therefore, such virus diseases can only be controlled through preventative means. As in the direct effects F. occidentalis have on crops, the different feeding behaviour of adult males and females is important in the transmission of viruses. Adult male F. occidentalis have a higher vector efficiency in transmitting TSWV than females (van de Wetering et al., 1999b). Viruses are obligate parasites, and TSWV needs to be injected into a plant cell, in saliva through feeding activity, which is still viable after feeding. Unlike adult females, whose intensive feeding activity results in plant cells being destroyed, males have a greater probing activity that is less likely to destroy the cells (van de Wetering et al., 1998). This results in TSWV being injected into viable plant cells, causing differential vector competency (van de Wetering et al., 1999a). 1.3.4 Control: Integrated Pest Management IPM may be defined as “a pest management system that, in the socio-economic context of farming systems, the associated environment and the population dynamics of the pest species, utilises all appropriate techniques in a compatible manner to maintain pest population levels below those causing the economic damage” (Dent, 1995). IPM is composed of four different control components: biological, chemical, physical and cultural control. 6 T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 1.3.4.1 Biological control Biological control is a crucial IPM component used to control F. occidentalis in UK glasshouse crops, especially cucumbers, and may be defined as “the use of a living organism (the beneficial) for the regulation of populations of another (the pest)” (Jacobson, 1997). The principal agents used in the biological control of F. occidentalis are predatory phytoseiid mites and heteropteran bugs (reviewed by Sabelis & Van Rijn, 1997). The principle predatory mite used is Neoseiulus (=Amblyseius) cucumeris (Oudemans) (figure 1.5), although 15 other Amblyseius spp. are reported as attacking F. occidentalis (Sabelis & Van Rijn, 1997). This mite only feeds succesfully on the egg and larva I stages of F. occidentalis (Gillespie & Ramey, 1988), as all other stages are capable of either escaping or defending themselves by jerking the abdomen (van der Hoeven & van Rijn, 1990), a response similar to that shown by T. tabaci to N. barkeri (=Amblyseius mckenziei) (Hughes) (Bakker & Sabelis, 1989). However, in the presence of another pest, the spider mite Tetranychus urticae Koch, thrips larvae move into the web produced by the spider mite to escape predation following detection of volatile cues associated with N. cucumeris (Pallini et al., 1998). This however results in reduced thrips developmental rate, due to interspecific competition for food between the thrips and the spider mites. N. cucumeris is most effective when used prophylactically and released onto the crop at the time of planting (Jacobson et al., 2001b). Using this strategy F. occidentalis populations can be almost entirely suppressed. If the release of the mite is delayed there is an increased risk of control failure, which may require the use of remedial chemical control that could disrupt biological controls in place for the control of other pest species, e.g. the hymenopteran parasitoid Encarsia formosa Gahan used to control the whitefly Trialeurodes vaporariorum (Westwood). Predatory heteropterans, such as the minute pirate bug Orius majusculus (Reuter), are occasionally used as a remedial treatment, but tend to be ineffective as they are slow to reproduce and oviposit into crop parts that are removed during normal crop maintenance, in cucumbers (Jacobson, 1993). Preventative release of Orius laevigatus (Fieber) on cucumber is ineffective as this species does not fully establish, but does on pepper (Chambers et al., 1993), possibly due to the presence of pollen as an alternative food source. Supplying sweet pepper pollen on cucumber for O. laevigautus serves as an effective alternative food source in the absence of thrips, but does not prevent searching for thrips when they are available (Hulshof & Jurchenko, 2000). This approach may allow the preventative use of Orius spp. against F. occidentalis in the 7 T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 future. However, the presence of additional pollen, which F. occidentalis feed on and improves their reproductive fitness (Trichilo & Leigh, 1988), may result in a pest population build-up, as well as reducing the predation rate of predatory mites, such as N. cucumeris (van Rijn & Sabelis, 1993). To a lesser extent, entomopathogenic fungi are also being used in IPM systems (reviewed by Butt & Brownbridge, 1997). Verticillium lecanii (Zimmermann) and Metarhizum anisopliae Zimmermann can cause 53% and 75% mortality of pupal stages, respectively, when larvae II are exposed to the fungal spores (Helyer et al., 1995). Newly emerged adults exposed to Paecilomyces fumosoroseus (Wize) suffer 58-90% mortality, within 4 d of exposure (Gindin et al., 1996) and Beauveria bassiana (Balsama) Vuillemin can reduce numbers of immature stages by 75% over a three week period on glasshouse cucumbers, without impairing control by N. cucumeris (Jacobson et al., 2001a). The compatibility of B. bassiana and N. cucumeris may therefore provide effective control of F. occidentalis without the need for remedial insecticidal treatment: prophylactic release of N. cucumeris would control F. occidentalis at acceptable levels and treatment with B. bassiana could be used if the F. occidentalis population exceeded the ability of the mites to control them, without affecting the mite population (Jacobson et al., 2001a). In the future it is likely that parasitic nematodes (reviewed by Loomans et al., 1997) will also be used in the regulation of F. occidentalis populations. The parasitic nematode species Heterorhabditis megidis Poinar, Steinernema feltiae (Filipjev) and S. carpocapsae Weiser caused 63%, 60% and 77% mortality of pupae in compost, compared to an untreated control mortality of 31% (Helyer et al., 1995), whilst H. bacteriophora (Poinar) reduced adult emergence to 40% of the control (Chyzik et al., 1996). It is likely that nematodes would only be of use in IPM when combined with predatory mites and bugs controlling the larval stages. 1.3.4.2 Chemical control There are several problems associated with the use of chemical insecticides to control F. occidentalis. The thigmotropic behaviour of thrips results in inefficient targeting of the insecticides and is likely to have a detrimental impact on biological control in use. There is also growing evidence of F. occidentalis resistance to existing insecticides, e.g.the pyrethroids permethrin and cyfluthrin, the carbamates methomyl and methiocarb, the organophosphates dimethoate andacephate, and the organochlorine DDT (Brødsgaard, 8 T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 1994; Immaraju et al., 1992; Robb et al., 1995; Jensen, 2000). This situation is compounded by insecticides being removed from commercial use on food crops and the fact that producers of insecticides are not registering new products for use against F. occidentalis. Supermarkets have considerable power in determining the usage of insecticides on food crops, and in recent years, have been demanding that the use of insecticides is reduced. 1.3.4.3 Physical control The main physical control measures against F. occidentalis in glasshouses are the use of traps and screens. Traps have five roles: (1) to detect thrips; (2) to enable comparison of F. occidentalis activity before and after treatments; (3) to trap thrips as they enter glasshouses through vents; (4) to ‘mop up’ thrips in empty glasshouses and (5) to reduce the numbers of thrips in localised ‘hot spots’ of activity within the crop (Jacobson, 1997). The colour and position of traps within the crop affect trap effectiveness. Blue, yellow, violet and white - UV coloured traps have been reported as being more attractive than white +UV or black traps (Gillespie & Vernon, 1990; Vernon & Gillespie, 1990), although Brødsgaard (1989) found that F. occidentalis were attracted to a specific shade of blue more than other shades of blue, yellow or white traps. Females are relatively more attracted to blue traps than males compared to yellow traps (Gillespie & Vernon, 1990). Traps placed against a contrasting colour are more attractive than traps placed against the same colour (Vernon & Gillespie, 1995) and colours that reflect UV are not attractive to flower thrips (Kirk, 1984). In cucumber crops, significantly more F. occidentalis are caught at a height of 2.4 m, compared to heights of 0.6 m and 1.2 m in a crop 2.1 m high (Gillespie & Vernon, 1990). The attractiveness of traps to thrips can be further enhanced by baiting the traps with attractive chemicals (§1.4.1), e.g. p-anisaldehyde (Brødsgaard, 1990; Teulon et al., 1993). Screens may be used to prevent thrips gaining access to the glasshouse via the vents, although the extremely small mesh size required for thrips exclusion may impede airflow. Mesh, with an aperture size that does not exclude thrips, coated with aluminium, which reflects UV, does discourage thrips from getting into glasshouses, as does aluminium tape placed around vents (McIntyre et al., 1996), and can reduce the ingress of thrips by as much as 55%. The use of materials that reflect or transmit UV to reduce the number of thrips on crops (reviewed by Terry, 1997; Antignus, 2000) also extends to UV reflective 9 T h e a larm p h er om on e o f th e wes te r n flo wer th r ips1 mulches between crop rows although the effectiveness is reduced as the crop canopy grows and shades the mulch. 1.3.4.4 Cultural control Cultural control is based on techniques that modify the agroecosystem to the disadvantage of the pest species. This may be achieved through modification of the host, e.g. using crop cultivars that are resistant to the pest (§3.1.1.2) and reduce developmental rate, or the environment, e.g. nursery hygiene, intercropping, the use of trap crops, supplementary lighting and manipulation of temperature and humidity. Nursery hygiene is extremely important in controlling F. occidentalis in glasshouses. Weeds outside the glasshouse and crop residue within the glasshouse may serve as a reservoir of F. occidentalis. This is especially important if replanting occurs during the growing season, as is the practice in UK cucumber crops. The replanted crops are vulnerable to F. occidentalis attack and biological controls applied to the plants may not be able to control a large, sudden invasion of F. occidentalis. This would result in remedial chemical treatments being used to control the invasion, causing collapse of biological control. Intercropping leek with clover results in a large reduction in the number of adult T. tabaci present on the intercrop compared with the monocrop (Theunissen & Schelling, 1993). This may be due to clover altering the host-plant quality resulting in an increased emigration rate of T. tabaci from the host-plant (den Belder et al., 2000). However, intercropping chrysanthemums with clover to combat F. occidentalis has resulted in increased thrips damage, possibly as a result of changes in the reaction of the chrysanthemum to thrips mediated by the clover (den Belder et al., 1999). At present, the use of intercropping in glasshouse cucumber is unlikely. Trap crops are used to selectively attract pests away from the main crop. The trap crop is then selectively treated, with either insecticides or biological control agents. The selective application of treatment to trap crops, such as the Verbena cultivar ‘Sissinghurst Pink’ (Bennison et al., 1999), not only reduces the expense of treatment to the grower, but also avoids inducing the breakdown of biological control throughout the crop by the application of insecticides. Biological control agents, such as Orius insidiosus (Say), enter reproductive diapause under the short photoperiod conditions that occur during the temperate winter months. The 10