Gilpinia hercyniae (spruce sawfly)

Gilpinia hercyniae is an invasive pest species; good examples are its introduction and rapid spread in Britain and North America. The thelytokous parthenogenesis permits establishment of the insect from a single female provided it reaches a stand containing spruce.

Diprionidae is a small family with about 125 species in 11 genera (Smith, 1993; Viitasaari, 2002). Gilpinia hercyniae (Hartig), the European spruce sawfly, was for a long time considered synonymous with the closely related spruce sawfly, Gilpinia polytoma (Hartig), until Reeks (1941) proved its validity.Adams and Entwistle (1981) described the confused history of the taxonomy of these two species roughly as follows. Both species were described by Hartig, G. polytoma as Lophyrus polytomus (Hartig, 1834, 1837), and G. hercyniae as Lophyrus hercyniae (Hartig, 1837). Rohwer (1910) pointed out that Lophyrus is preoccupied and that the proper genus is Diprion. Enslin (1912) perpetuated the genus Lophyrus and relegated L. hercyniae as a synonym of L. polytomus. Escherich (1913) still recognized polytomus and hercyniae as distinct and continued to place them in Lophyrus. Smith (1938) discovered cytological and biological differences between Diprion polytomum in Canada and in Europe.In the 1930s this spruce sawfly species had suddenly started to spread in northeastern North America and to cause serious damage in spruce forests. Benson (1939) proposed a new genus Gilpinia in which he placed polytoma, and apparently did not recognize hercyniae as distinct. Smith (1941) recognized two European forms, one of which was cytologically identical with the Canadian form of D. polytoma. The separate identities of Gilpinia polytoma and G. hercyniae were re-established by Reeks (1941), and confirmed by Forster (1949), and the latter species alone was shown to be present in North America (Balch et al., 1941). The two species had different chromosome numbers: 2n = 14 in G. hercyniae, and 2n = 12 in G. polytoma (Reeks, 1941; Smith, 1941).In 1942 the name Gilpinia hercyniae began to come into use in the literature (e.g. Balch, 1942; Morris, 1942; Prebble, 1943). However, some authors persisted in the use of Diprion whilst others began to revert from Gilpinia to Diprion (c.f. Balch, 1942, 1958; Reeks, 1952, 1953, 1963). Smith (1974) published a checklist of Diprionidae of the world, in which he listed hercyniae and polytoma under Gilpinia. Nevertheless some authors in the 1970s continued to perpetuate Diprion hercyniae. Due to the taxonomic confusion, G. hercyniae was, prior to Reeks (1941), often erroneously referred to in the literature by the name Diprion polytomum. The occurrence of the three common European species of spruce sawflies, Gilpinia hercyniae (Hartig), G. polytoma (Hartig), and G. abieticola (Dalla Torre) has often been dealt with more or less collectively in the literature (Thalenhorst, 1955, 1960; Pschorn-Walcher, 1974, 1982).See Adams and Entwistle (1981) for an annotated bibliography of Gilpinia hercyniae, covering the period 1834 to December 1979.

EggsEggs are pale green, elongate oval, about 1.9 mm long and 0.5 mm wide. Eggs are mainly laid singly into the edge of the one-year old needles. The female, using her saw-like ovipositor, makes a longitudinal slit, egg-pocket, in the middle or apical half of the needle and injects the egg. Freshly laid eggs are hard to detect, but a needle containing a viable, developing egg has a distinct bulge and the unhatched egg can often be seen protruding from the oviposition slit. Often the egg-pocket turns pale brown in colour (Scheidter, 1934; Billany, 1978).LarvaeThere are five feeding larval instars. The body colour of the first four instars is mostly green, but the fourth instar begins to show white stripes. L5 instar, 15-20 mm long, is green with five white stripes running the full length of the body; one mid-dorsal, one dorso-lateral, and one lateral; also a blackish dorso-lateral and a broken pedal stripe exist. The head capsule is mostly blackish in the first instar, becoming browner in second and succeeding instars. Spiracles are light brown. The fully grown larva moults to the final non-feeding prepupal or prespinning larva with a greenish brown head capsule and body without white stripes. It pupates by spinning a cocoon.Green forms of larvae of G. hercyniae and G. polytoma are often regarded as indistinguishable (Reeks, 1941; Thalenhorst, 1955, 1960; Verzhutskii, 1973). Vehrke (1961) described characteristic differences in the dark frontal figures on the head capsules, which enable in most cases the separation of the larvae of these two species. Due to a considerable degree of variation in the pigmentation pattern of the head, especially in G. hercyniae, an exact identification is sometimes not possible (Thalenhorst, 1960; Pschorn-Walcher, 1974). Keys to larvae are given by Benes and Kristek (1979) and Wong and Szlabey (1986). Besides the green larval form, G. polytoma has a rarer form which is ventrally salmon-red (Reeks, 1941; Smith, 1941; Sellers, 1942).The neuroendocrine organs of G. hercyniae larva were described by Hinks (1973).CocoonThe cocoon is cylindrical to spindle-shaped with rounded ends, tough, and finely textured. It is reddish brown, about 8-9 mm long and 4 mm wide (Billany, 1978, Plate 3), and has a weight of 70-75 mg (Otto, 1991). Cocoons are spun in the litter.AdultsReeks (1941) translated Hartig's original descriptions of G. polytoma (Lophyrus polytomus) and G. hercyniae (L. hercyniae), and he also redescribed both of the species. Only differences in genitalia of males and females allowed unambiguous separation of the species. A new character, maculation of the frons of adults, was presented by Benes and Kristek (1979), and Goulet (1981). In G. polytoma the surface ventrad to median ocellus with sharply outlined shallow fovea; fovea elongate, triangular, testaceous, and smooth; in G. hercyniae the surface ventrad to median ocellus without shallow fovea; fovea rugulose, sparsely punctate, and black or piceous.Males: 4-8 mm, black, abdomen ventrally yellowish to ferruginous, darker on sides. Labrum and clypeus generally brownish or luteous, sometimes whitish. Angles of pronotum yellow. Antennae black with 24-28 segments, bipectinate. Penis valve pedate and strongly sclerotized at apex. In G. polytoma labrum and clypeus mostly whitish, and the penis valve spatulate with distal third membranous.Females: 6-9 mm, black with yellow markings. Head mostly yellow; a blackish band between compound eyes enclosing ocelli and extending to antennal fovea. Labrum brownish, compound eyes black. Antennae with 21-22 segments, flagellum dark, serrate. Lancet with 11-12 annuli (in G. polytoma 9-10 annuli). Goulet (1981) has described characteristic differences between females of G. hercyniae and G. polytoma in the form of antennomere 3, in the distance between lateral ocellus and nearest inner margin of eye, and in the height of eye/ocellar-ocular distance.Males of G. hercyniae are extremely rare (1 male to 1000-1200 females). A gynandromorphic adult, exhibiting characteristics of both sexes, has been described by Niklas (1962a).

Due to the taxonomic confusion between G. hercyniae and G. polytoma, many records prior to 1941 cannot be reliably identified to the species (e.g. Sellers, 1942). G. hercyniae is native and widely distributed throughout central and northern Europe to Siberia, Mongolia, Korea and Japan (Smith, 1974; Pschorn-Walcher, 1982; Zhelochovtsev, 1988, 1994; Liston, 1995). It was established as an introduced pest species in Britain where it was first found in 1906 (Benson, 1933, 1951). It occurs in England and Scotland but not in Northern Ireland or the Republic of Ireland (Billany and Brown, 1977; Billany, 1978; Bevan, 1987, Liston, 1995).Gilpinia polytoma is native and distributed in Central and Northern Europe, in Siberia to the Pacific coast, in Japan and in Pakistan. The species has not been recorded in Britain nor in America (Smith, 1971, 1974; Pschorn-Walcher, 1982; Mohyuddin et al., 1984; Zhelochovtsev, 1988, 1994; Liston, 1995).Also see CABI/EPPO (1998).

G. hercyniae, the European spruce sawfly, was accidentally introduced into North America, and was first found in Ontario, Canada, in 1922 and in New Hampshire, USA, in 1929. The first massive outbreak started in the Gaspe Peninsula, Quebec, in 1930, and the dispersal rate in Canada was 30 miles per year (Balch, 1939; Reeks, 1963; Buckner, 1966). Prior to 1941, the species found in North America was referred to under the name Diprion polytomum (Hartig) (Reeks, 1941; Smith, 1979).

The EEC Plant Health Directive, laying down measures to protect the European Economic Community from entry of foreign pests and diseases, was published for the first time in 1977 (Phillips, 1978). The "protected zones" (PZ) against G. hercyniae were Greece, Ireland, and Northern Ireland, the Isle of Man and Jersey. The same protected zones against G. hercyniae are also included in the Council Directive 2000/29/EC, of 8 May 2000, on protective measures against the introduction into the Community of organisms harmful to plants or plant products and against their spread within the Community.

Since the adults are strong fliers, and sometimes reach considerable altitudes, they must often travel many miles with the assistance of air currents. The thelytokous parthenogenesis permits establishment of the insect from a single female provided it reaches a stand containing spruce. Large numbers of adults have been observed on the tops of mountains above the timber line, and a great many adults drop into lakes and rivers (Balch, 1939).G. hercyniae larvae normally pupate (spin cocoons) in the litter and soil, occasionally also in the foliage or branches. During outbreaks there is a risk that larvae accidentally pupate in containers, vehicles, camping equipment etc. in the forest, and thus get transported into new areas.

Gilpinia hercyniae attacks, to a varying degree, practically all species of spruce (Picea spp.). In Europe, Picea abies (P. excelsa) is the main host. According to Billany (1978), all age classes and species of spruce are susceptible to attack, and no significant host preference has been noticed with respect to provenance, aspect or elevation. All species of spruce native or grown in eastern North America are hosts but some preference is shown for P. glauca in oviposition and development (Balch, 1939; McGugan and Coppel, 1962).Picea omorika is more resistant to G. hercyniae than P. abies (Dominik, 1989). In feeding experiments, larvae rejected P. omorika needles (Ohnesorge and Serafimovski, 1961). Feeding on Abies balsamea (Browne, 1968) and on A. alba (Schedl, 1982) has also been recorded, but evidently only in the absence of the natural host.

Solitary larvae feed from June to September/October on fully developed, mainly 1-year-old or older, foliage of spruces. The small early stage larvae are difficult to spot owing to their colour, which closely matches the foliage, and their tendency to lie along the needles. Late stage larvae are more obvious, being bright green with white stripes running the full length of the body.Larvae prefer the lower branches of trees and the edges of stands (Ohnesorge and Thalenhorst, 1956; Pschorn-Walcher, 1974). Even in light infestations, feeding is heavier in the lower part of the crown, and a severe outbreak gives a greyish-brown cast to the stand (MacAloney, 1936). The sawfly tends to kill the crown from the bottom upwards (Reeks and Barter, 1951). During heavy outbreaks, trees may gradually lose all of their old and part of their new foliage, with only a green tuft being left at the top. In Canada, heavily defoliated and dying trees were attacked by the eastern spruce barkbeetle (Dendroctonus piceaperda) and by secondary barkbeetles (Dryocoetes, Ips, Polygraphus) (Balch, 1941; Reeks and Barter, 1951).According to Bevan and Davies (1971), the first signs of defoliation on Picea sitchensis appear in the upper part of the crown, and become noticeable when the current year's shoots show up as a reddish yellow colour. The early signs of rising population can be seen in the top four whorls of the crown (Billany, 1978).Perhaps the best indication of the presence of larvae is the accumulation of large pellets of bright green frass approximately 2 x 1 mm diameter which lodge in the foliage or fall to the forest floor (Morris, 1942b; Billany, 1978). The green colour of the frass turns reddish brown after exposure to weather for a number of days. During heavy outbreaks, the frass may cover the ground to a depth of half an inch (1.25 cm) in some places (Dirks, 1944). A frass key for certain spruce defoliators, including G. hercyniae, has been published by Morris (1942b), and for certain defoliators of forest trees by Hodson and Brooks (1956).

Spruce sawfly species are morphologically so similar that an exact identification in the field is often difficult. Therefore samples of larvae, adults or cocoons should always be taken for later determination.According to Pschorn-Walcher (1974), larvae of Nematinae are often difficult to separate from those of Gilpinia spp. without proper inspection under the microscope. Almost all the spruce Nematinae in Central Europe complete feeding in July or early August, so that in later collections (after mid-August) virtually only larvae of Gilpinia spp. are obtained. The sorting, handling and rearing of this rather "pure" material is then much easier and less time-consuming (Pschorn-Walcher, 1974). Keys to spruce sawfly larvae have been given by Lindquist and Miller (1971) and Benes and Kristek (1979).

The life history and habits of G. hercyniae were investigated intensively after the species was introduced and established as a major pest in North America. The literature has been reviewed by Smith (1979), Adams and Entwistle (1981) and Pschorn-Walcher (1982). G. hercyniae is obligatory parthenogenetic and thelytokous, i.e. females produce females and only very rarely males. Sex ratio is 1 male to 1000-1200 females. G. polytoma is facultatively parthenogenetic and arrhenotokous, i.e. unfertilized females produce males. Sex ratio is about 1:1. The number of generations of G. hercyniae per year varies, depending upon latitude and climatic conditions. In the most northerly latitudes there is only one generation per year, in the middle region two are possible, and in the southern extremity, three generations are produced. Adults of univoltine populations fly mainly in May-June (July), those of the bivoltine populations in May-June and July-August, and the flight periods in the south are in April-May, June-July and August-September, respectively. According to Billany (1978) adult sawflies may be found throughout the summer until late September. There is great temporal variation in the emergence and larval periods between years depending on weather conditions, and the generations may overlap. The eggs are laid singly in slits in old needles and hatch in about 10 days. Soft walled shade needles are best suited for oviposition (Bombosh and Ramakers, 1976), but Billany et al. (1978) found a preference for 1-year-old needles at the top of the crown of P. sitchensis. A single female may lay 30-60 eggs (Thalenhorst, 1968). The larvae live solitarily and in about 40 days, pass through five feeding stages, and one non-feeding stage. The effect of temperature on the development of the eggs and/or larvae has been studied by Thalenhorst (1973), and Otto (1991). Larvae prefer old foliage for feeding. Jensen (1988) reported 100% mortality of larvae when fed with newly-flushed foliage. The late season larvae can consume fully-matured new foliage (Balch, 1939). About 80 per cent of the feeding is done by the fifth larval instar (Reeks and Barter, 1951). The pre-spinning or prepupal larva spins a cocoon in the litter layer beneath the tree. G. polytoma spins cocoons mainly in the foliage or branches. The eonymphs in the cocoon may enter diapause or produce adults within a fortnight. The winter is passed in the eonymphal or pronymphal stage in the cocoon. A portion of the overwintering population may remain in diapause (prolonged) for two or more, sometimes even five, seasons. Diapause has a genetic base but temperature, day length (photoperiod), and moisture also regulate it (Prebble 1941; Niklas, 1962; Bombosch and Ramakers, 1976). The rearing of larvae under long day conditions (17 h) at 25°C inhibits the induction of eonymphal diapause, and even nine generations of certain strains have been reared annually in the laboratory (Bird, 1961). The life history and habits of G. hercyniae make it an ideal insect for laboratory research; it reproduces parthenogenetically, it is thelytokous, the eggs are deposited singly, larvae feed on old growth spruce foliage available throughout the year, and several generations may be reared annually. In the University of Göttingen, Germany, G. hercyniae has, since 1970, been the test insect for a long series of investigations on the nutritional quality of spruce needles for forest insect pests and on the interaction between the host tree and the pest (Bombosch, 1972; Lunderstädt and Claus, 1972; Lunderstädt, 1973, 1977, 1981a,b,c, 1983; Schwarz, 1973; Dinish, 1974; Lunderstädt and Hoppe, 1975; Lunderstädt et al., 1975, 1976/77; Dohrenbusch, 1978; Müller, 1979; Schopf, 1979, 1980; 1981, 1982, 1983a,b, 1986a,b,c, 1989; Lunderstädt and Reymers, 1980; Müller et al., 1981; Schopf et al., 1982; Lunderstädt and Ahlers, 1983, 1984, 1985; Kulcke, 1983; Lunderstädt and Küster, 1985; Spiller and Lunderstädt, 1988; Sabsch and Schopf, 1990; Schopf and Hartl, 1997). G. hercyniae is, despite of its wide distribution, a species that rarely occurs in outbreaks (Larsson et al., 1993). The main factors keeping the populations at endemic level are a nuclear polyhedrosis virus disease, parasitoids, and predators. The few economic level outbreaks known (Eastern North America, UK), have all occurred in cases where the sawfly has been introduced into new territories without its natural enemies. The taxonomic status, morphology, bionomics and ecology of G. hercyniae have been studied by Borowski (2004a,b,c).

G. hercyniae is attacked by several hymenopterous and dipterous parasitoids; many predators including ants, bugs, beetles, spiders, small mammals, birds; and a species-specific nuclear polyhedrosis virus (GhNPV).G. hercyniae larvae possess a pair of esophageal diverticulae with resinous contents. In response to harrasment, larvae regurgitate a droplet of the resinous liquid. This behaviour, typical for the whole family of Diprionidae (Codella and Raffa, 1993), has been shown to repel predators and parasitoids. The liquid acts also as an oviposition deterrent for conspecific females (Weitzel and Hilker, 1993).There are a large number of parasitoid records in the literature. For various reasons, however, a relatively high proportion of these records, particularly the older ones, are unreliable. The main sources of error have been the taxonomic confusion between G. hercyniae and G. polytoma prior to 1941, the mass-collection and mass-rearing of diprionid larvae and cocoons, inadequate sample-size and faulty timing of sampling, and false identification of the parasitoids or the hosts (see Morris et al., 1937; Reeks, 1952; Oehlke, 1965; Pschorn-Walcher, 1965, 1974). Morris et al. (1937) found 31 species of hymenopterous and dipterous parasitoids of Gilpinia polytoma (Diprion poytomum) in the material collected in 1932-1936 in Europe, and a total of 64 species of primary and secondary parasitoids were reared from this host in the material collected in 1935-1939 in Europe for shipment to Canada (Finlayson and Finlayson, 1958).Egg parasitoids are not known from G. hercyniae. According to Vikberg (1985), a chalcid tetracampid, Dipriocampe elongata (Erdös), might be an egg parasitoid of G. hercyniae since he swept both of these species on the same Picea abies trees, and the parasitoid was known to be associated with P. abies in Hungary and Czechoslovakia.The most important parasitoids of G. hercyniae in Europe are the larval parasitoids Olesicampe (Holocremnus) ratzeburgi (O. macellator), Lamachus marginatus (L. ophthalmicus, L. coalitorius), Exenterus tricolor, E. confusus, E. vellicatus, E. amictorius (Ichneumonidae), and Drino bohemica (Tachinidae), and the cocoon parasitoids Pleolophus basizonus, Agrothereutes spp. (Ichneumonidae) and Dahlbominus fuscipennis, Monodontomerus dentipes, Mesopolobus (Amblymerus) subfumatus (Chalcidoidea). Total parasitism usually varies over about 10-25% (Pschorn-Walcher, 1974).When the massive outbreak of G. hercyniae in Canada in 1930 was detected, a virtual absence of attack by native parasitoids was found. Of the 276,000 cocoons and mature larvae, collected throughout the whole of the known range of the sawfly in Canada during 7 years, only less than 0.02% were successfully parasitized by native species (Balch, 1939). An extensive biological control programme was initiated in Europe in 1932 with a view to introduce natural enemies into Canada in order to bring the pest under control. The results of this successful programme are covered in the Control section. According to Finlayson (1960), 32 species are known in the literature as parasitoids of G. hercyniae in Canada.Common insect predators preying on larvae and occasionally on adults of G. hercyniae are the pentatomid heteropterans, Podisus serieventris (Reeks, 1938) and P. maculiventris (MacAloney, 1936).According to Balch and Bird (1944) some sawfly adults are destroyed by birds, but the number does not seem to be large. Entwistle et al. (1977a,b, 1978) reported that at least eight species of birds (wren, goldcrest, robin, redwing, song thrush, coal tit, blue tit, and starling) eat larvae of G. hercyniae in the field. They also found that 16 species of birds belonging to six families dispersed infective NPV in their faeces. In Saxony, Escherich and Baer (1913) found cocoons of Lophyrus hercyniae in the stomachs of sparrows and tits during winter, but the species was almost certainly Gilpinia polytoma.Adults of coleopteran Carabidae readily attack sawfly cocoons in the litter (Morris, 1951). Larval Elateridae have been reported to destroy 1-2% (Morris, 1951) and 10-18% (Martineau, 1963) of the cocoons of G. hercyniae in the litter.Small mammal predators, shrews, voles and mice (Soricidae, Microtidae), are very important factors affecting the spruce sawfly cocoons. In general, the portion of cocoons destroyed by small mammals has been reported to be 30-50% (Morris, 1942a, 1949a, 1951; Martineau, 1963; Neilson and Morris, 1964; Billany, 1978).Larvae of G. hercyniae are infected by a lethal species-specific nuclear polyhedrosis virus disease (GhNPV). The disease is caused by Borrelinavirus hercyniae (B. gilpiniae), which belongs to baculoviruses (Tinsley et al., 1980; Cunningham and Entwistle, 1981). It occurs widely in nature and is one of the main factors keeping the populations of G. hercyniae at low level. The disease is not known from Gilpinia polytoma. The nature of the disease and symptoms of infection were described by Balch and Bird (1944). The literature on the disease and its role in regulating host populations have been reviewed by Adams and Entwistle (1981), and Cunningham and Entwistle (1981). Larvae become infected either by ingesting the polyhedral inclusion bodies (PIB) with the food, or via infected parent (Bird, 1961). The period of infection to death is 6-11 days. Parasitoids, predators and scavengers may act as transmission agents by straight contamination, or via their infectious faeces (Entwistle et al., 1977, 1978). GhNPV has been an effective factor in the biological control programmes against G. hercyniae, see Control section.

Gilpinia hercyniae has no pest status in its native distribution range. In Central Europe some minor outbreaks have been caused by spruce sawfly species not exactly identified (Thalenhorst, 1960; Pschorn-Walcher 1974, 1982). Infestations of economic level damage have occurred only in territories and countries into which G. hercyniae was accidentally introduced and established without its natural enemies. Well known examples are the outbreaks in Britain and in northeastern North America.In Britain, a minor outbreak, which declined within a few years, was noted on Picea omorika in 1950 (Billany, 1978). The outbreak history of G. hercyniae in Britain in 1970-1977 is well documented (Bevan and Davies, 1971, 1972; Brown, 1973; Brown and Billany, 1974; Bevan, 1975; Billany and Brown, 1975; Billany, 1976, 1977, 1978; Carter et al., 1976; Billany et al., 1977). Defoliation and damage to Picea sitchensis and P. abies were widespread and locally serious, but the trees were not killed. The effect of five years defoliation on the height increment of 16-year-old Sitka spruce brought about a 24% reduction. No reduction in radial increment was detectable. In many cases of top-kill of P. abies, side shoots took over and recovery was good though stem distortion through multiple leaders was not uncommon (Billany, 1978).The huge literature on the introduction of G. hercyniae into North America in the 1930s and the large biological control programme initiated in 1932 to stop the destructive spread of the new pest in the northeastern parts of Canada and USA, have been thoroughly reviewed by McGugan and Coppel (1962), Neilson et al. (1971), Reeks and Cameron (1971), Clausen (1978), Adams and Entwistle (1981), Hulme and Green (1984), and Magasi and Syme (1984).The outbreak started in the Gaspe Peninsula, Quebec, Canada, in 1930 and over 15 years spread westwards to Ontario, Canada and southwest to the New England states, USA (Balch and Simpson, 1932; Balch et al., 1934; MacAloney, 1936; Dowden, 1939; Balch, 1937, 1939, 1940, 1941; Balch and Bird, 1944). The most severely defoliated were Picea glauca, P. mariana and P. rubens. According to Reeks and Barter (1951) and Kulman (1971) loss of 20-50% of the old foliage caused measurable growth losses. Successive defoliations for 13, 7, and 4 years resulted in increment losses for 17, 12, and 10 years. In Picea glauca, 4-7 years of 90% old and 50% new foliage loss caused little mortality, but trees that lost 95% old and 75% new foliage frequently died. P. mariana died from lower rates of defoliation than did P. glauca.Severe mortality was confined to Gaspe, where the outbreak lasted 8 to 15 years. By 1940 it was estimated that 66% of the spruce volume had been destroyed and the total mortality in the Gaspe was estimated at over 40 million m³ of spruce (Reeks and Barter, 1951). Additional loss resulted from reduced growth on surviving trees and associated windfall. According to Riley and Skolko (1942), spruces killed by sawfly deteriorate faster than bark-beetle killed trees, because the bark remains intact keeping moisture content high.Reeks and Cameron (1971) have evaluated the economic losses caused by the outbreak in Canada, and the benefits produced by the biological control programme. The collapse of the outbreak by 1945 was mainly due to the effectively spread NPV disease and the action of introduced and native parasitoids. Populations have generally remained low since the early 1940s, and G. hercyniae could be relegated, at least in terms of economic importance, to the status of an unimportant insect (Neilson et al., 1971; Magasi and Syme, 1984).Williams et al. (2003) demonstrated that the needle trace method (NTM) can be applied successfully to Sitka spruce (Picea sitchensis) to quantify defoliation caused by an outbreak of G. hercyniae that occurred many years previously, and that the technique can provide data on needle loss that is valuable for interpreting reductions in tree growth.

During heavy outbreaks of G. hercyniae the mortality of the more susceptible Picea species may lead to their decrease and substitution by less susceptible Abies species in the stands (Reeks and Barter, 1951; Reeks and Cameron, 1971).

Eggs and the early stage larvae of G. hercyniae are difficult to detect. Larvae and the occasional adult female can be collected by beating the foliage with a stick over a beating tray (Thalenhorst, 1960; Neilson and Morris, 1964; Pschorn-Walcher, 1974). A good indication of the presence of larvae is the accumulation of pellets of bright green frass in the foliage or on the forest floor (Morris, 1942b; Billany, 1978). Frass drop measurements can be used also for determining population trends (Morris, 1949b).G. hercyniae overwinter in a cocoon in the forest litter. In plantations on ploughed land, the sawfly shows a distinct preference for overwintering on the ploughed ridge, but despite this tendency to aggregate, cocoons are very difficult to find at low population levels (Billany, 1978). Various methods of cocoon sampling in population studies have been tested by Prebble (1943), and a non-destructive, wet technique to extract sawfly cocoons from a variety of soil and surface litter types has been described by Jukes (1977).Quite considerable defoliation and increasing populations of the sawfly can go undetected for several seasons in dense stands of spruce. This is due to the solitary feeding and the preference of larvae for older needles. Spruces retain their needles for several years and the current year's growth also tends to mask defoliation in light to moderate attacks (Billany, 1978). So, the large and destructive outbreak in Canada, when first discovered in the Gaspe Peninsula in 1930, had been in progress for several years (Balch, 1939; Morris, 1958).

Biological Control

Larvae of G. hercyniae were not susceptible to Bacillus sotto toxin, probably because the toxin is not soluble in the gut due to the comparatively low pH of 8.5 (Angus, 1956). In laboratory tests Bacillus cereus was pathogenic (36-46% mortality) to larvae of G. hercyniae, but there was no suggestion that the bacterium could be used as a control agent (Heimpel, 1961). Smirnoff and Berlinguet (1966) found in some commercial preparations of Bacillus thuringiensis var. thuringiensis a labile exotoxin which caused a 100% mortality of larvae of G. hercyniae and many other sawfly species.

The most effective biological control agents against G. hercyniae are the nuclear polyhedrosis virus (GhNPV) and the parasitoids.

In 1932 an extensive biological control programme was initiated in Europe with a view to introduce natural enemies into Canada in order to bring the European spruce sawfly under control. This campaign was successful and by 1938 the sawfly population began to decline and continued to do so until, by about 1945, the outbreak in eastern North America had largely subsided (McGugan and Coppel, 1962; Neilson et al., 1971; Reeks and Cameron 1971; Pschorn-Walcher, 1982; Hulme and Green, 1984; Magasi and Syme, 1984).

In 1933-1951 over 890 million individuals of 27 parasitoid species, introduced and reared, were released against G. hercyniae in North America. Eight parasitoid species became established. Dahlbominus fuscipennis, Exenterus confusus and E. amictorius were effective mainly at high host densities whereas Exenterus vellicatus and Drino bohemica proved to be effective also at low host population levels. The main control agent, however, was the virus, which was accidentally introduced, most probably with parasitoid material, from Europe. In Canada the disease was discovered in larval laboratory cultures in 1936 and in the field populations in 1938. It spread rapidly and was also disseminated artificially, and by 1942 it was distributed throughout the greater part of the range of the sawfly (Dowden, 1940; Balch and Bird, 1944; Bird, 1961; Bird and Burk, 1961; Reeks, 1963). Biological control attempts against G. hercyniae using parasitoids and NPV virus in Canada were evaluated to have given good control for 1910-58, and complete control for 1958-68 and for 1969-80 (Hulme and Green, 1984). This control programme and the attendant studies must rank as one of the best documented and most successful attempts to control and regulate an insect pest in the history of biological control (Neilson et al., 1971).

The biological control programme for the importation of parasitoids from Europe to Britain, initiated in 1973, was less successful. Some recoveries of the released parasitoid species were made, but establishment of effective parasitoid populations were not recorded (Pschorn-Walcher, 1974; Greathead and Pschorn-Walcher, 1976; Billany, 1978; Billany et al., 1983). The virus arrived in Britain in 1972, probably carried by birds from Europe. It spread rapidly and controlled G. hercyniae more or less completely by 1977. The population of the pest has remained at a low level since then (Entwistle et al.,1977; Bevan, 1984, 1987).

G. hercyniae females lay their eggs singly and therefore are capable of starting many foci for virus infection. This might well account for the phenomenal spread of the GhNPV and for the very low population levels compared to the NsNPV of Neodiprion sertifer which lays eggs in clusters (Bird and Elgee, 1957; Bird, 1961; Bird and Burk, 1961). The patterns of spatial growth of small epicentres of GhNPV were studied in spruce forests in Wales by Entwistle et al. (1983). According to Dwyer (1994), the spatial spread of GhNPV can be explained with a very simple model of host movement, without recourse to complicated mechanisms of dispersal.

In the field, populations of G. hercyniae GhNPV maintained a very high virulence for over twenty years (Bird and Burk, 1961). Polyhedra of G. hercyniae lose virulence during storage, complete inactivation being reached after 12 years at 4.5°C (Neilson and Elgee, 1960).

Pheromonal Control

G. hercyniae reproduces parthenogenetically, and males are extremely rare. Mating instinct appears to be rudimentary, and only a few cases of attempted mating have been observed (Balch, 1939; Smith, 1941). Neilson and Morris (1964) state that mating has never been reported. The sexual pheromone of G. hercyniae has not been investigated (Anderbrant, 1999). So, monitoring and control methods based on sexual pheromones are not available.

Silvicultural Methods

G. hercyniae is a primary pest attacking quite healthy trees. Larsson and Björkman (1993) reported that larvae did not survive or grow any better on drought stressed trees (Picea abies) than on control trees. Bombosch and Lunderstädt (1976) studied the nutritional quality variations of spruce stands for G. hercyniae, and commented on the possible importance of this in forest management. Avtzis and Bombosch (1993) studied the effect of fertilizer application, pruning and thinning of P. abies on the quality of its foliage as food for the sawfly. Larvae developed better on vigorous trees with good needles, long crowns and little root competition than on untreated control trees with poor growth, small crowns and strong root competition. Applications of phosphate to the crops have been made in Britain because the heaviest outbreaks of G. hercyniae occurred on the phosphorus-deficient soils of mid-Wales (Bevan, 1987).

Chemical Control

Chemical insecticides have not been employed in control programmes against G. hercyniae. According to Balch et al. (1934) the sawfly is easily controlled by dusting, but economic conditions did not permit artificial control measures in Canada. Some small scale spraying of small decorative and shade trees was carried out by the landowners (Baldwin, 1939).

In Britain, a biological control programme against G. hercyniae was initiated in 1970s instead of chemical applications (Bevan and Davies, 1971; Brown and Billany, 1974; Pschorn-Walcher 1974; Billany, 1978).