Ecological impact of entomopathogenic nematodes used to control the large pine weevil, Hylobius abietis (Coleoptera: Curculionidae)

The large pine weevil (Hylobius abietis; Coleoptera: Curculionidae) is one of the most economically damaging pests in forestry across Northern Europe. Its larvae develop in coniferous tree stumps on clearfell sites and adult weevils feed on seedlings that are replanted on these sites, causing substantial mortality. As the drive towards achieving the sustainable management of natural resources increases, biological control agents are being considered as an alternative to chemical pesticides. Though such agents have great potential for contributing towards a reduction in the ecological impact of pest insect control, they must also be investigated for any potentially damaging effects they may have in whichever context they are being employed. Entomopathogenic nematodes (EPN; Steinernema carpocapsae and Heterorhabditis downesi) are currently being inundatively applied to tree stumps on coniferous clearfell sites in the UK and Ireland as biological control agents against the large pine weevil. The aim of the laboratory and field-based experiments and field sampling of clearfell sites that were carried out for this thesis was to investigate the risks associated with this use of EPN in this setting. To this end, the persistence of EPN in soil and bark after application to clearfell sites was recorded and attempts were made to investigate the possibility of hybridisation occurring between an exotic and a native Steinernema strain. Effects of EPN on some important non-target insects on clearfell sites, the service-providing wood decomposer Rhagium bifasciatum and the pine weevil parasitoid Bracon hylobii were also assessed in laboratory and field experiments. EPN presence in soil samples collected around tree stumps on clearfell sites five months, one year and two years after application of EPN decreased significantly from year one to year two. EPN presence on all sites sampled two years post EPN application was relatively low (less than 7 % positive soil samples). A similar trend was observed for EPN persistence under the bark of stumps. No EPN at all were detected two years after application on those sites treated on a small scale by manually applying nematodes. EPN persistence appeared to be better predicted by tree stump species (which in turn is indicative of the number of pine weevils developing within them) compared to soil type. Spread of nematodes was limited mostly to a 40 cm radius around stumps and little evidence of EPN presence in untreated areas was found on any of the investigated sites. While some spread of EPN to areas adjoining two of the sites occurred, there was no evidence of establishment off-site. (Chapter III). Attempts to distinguish between an indigenous and an exotic strain of S. feltiae and their hybrid using two molecular methods (Amplified Fragment Length Polymorphism or AFLP and Restriction Fragment Length Polymorphism or RFLP) were confounded by methodological obstacles. However, RFLP analysis of the strains did indicate a high rate of intraspecific and even intrastrain variation for S. feltiae. Using this method it was also possible to tentatively identify a field-isolated Steinernema sp. as Steinernema kraussei, though two other unidentified isolates could not be assigned to a particular species (Chapter IV). The saproxylic beetle community in deadwood other than tree stumps on Irish clearfell sites appears to be dominated by the longhorn beetle Rhagium bifasciatum (Coleoptera: Cerambycidae; Fabricius 1775). Both S. carpocapsae and H. downesi invaded decomposing logs and infected all stages of this species within logs when applied directly to them. However, infection only occurred to a large extent (> 50 % of insects) if the applied dose of EPN was half that applied to tree stumps (1.8 million IJs) to control the pine weevil and infection was also significantly lower when such logs were incubated in the field rather than the laboratory. R. bifasciatum individuals were infected in all regions of logs and baiting of wood samples revealed that IJs were penetrating logs to their very centre, indicating high mobility and infectivity of IJs within this substrate. S. carpocapsae emerged in numbers of 1,000 IJs or more per week for up to eight weeks after first emergence while H. downesi emergence had effectively ceased by the fifth week. A substantial number of live S. carpocapsae IJs remained in host cadavers even after eight weeks of emergence. Infection of R. bifasciatum was also recorded on three clearfell sites to which EPN had been applied around tree stumps previously (only a single infected larva found on one of the sites). Wood samples from some of these logs contained high numbers of EPN IJs (> 6,000), suggesting that EPN were recycling within logs. However, since infection of R. bifasciatum decreased significantly as the distance between a log and treated tree stumps increased it appears that the risk to this serviceproviding non-target insect can be minimised by being as accurate as possible when applying nematodes. (Chapter V). The ectoparasitoid wasp Bracon hylobii Ratz. (Hymenoptera: Braconidae) is closely associated with the large pine weevil and may therefore be at risk if EPN are applied against the latter (eggs laid on EPN-infected host larvae do not develop to adulthood). Female wasps did not parasitise dead hosts in standard trials (small bark patch [2.25 cm2]; host either freeze-killed or killed by EPN). However, naive wasps parasitised hosts that were moribund (i.e. those that died of EPN-infection within the 24 h period they were offered to wasps) significantly less frequently than either healthy control hosts or infected hosts that survived the 24 h trial period. Experienced wasps (one egg laying experience) parasitised all of these host types with similar frequency. Behavioural observations provided no evidence of wasps locating heavily infected or dead hosts and then rejecting them. Heavily infected hosts that died within the 24 h trial period during which they were offered to wasps tended to move less than their surviving counterparts and wasp oviposition was significantly correlated with host movement and host feeding during trials. NaЇve wasps were significantly less likely to parasitise hosts that had their mandibles glued shut than those that did not (this had no effect on parasitism by experienced wasps). Thus, it is suggested that vibrations caused by these host activities were the most important cue driving host location and/or host acceptance in B. hylobii. In continuous 30 minute observation of trials in which wasps were offered a host at the centre of a large bark patch (20.25 cm2), experienced wasps spent more time close to the host when it was moving and were also significantly more likely to approach the host when it was moving, indicating a vibrotaxic response. Moreover, they only were attracted to dead hosts (freeze-killed or killed by EPN) when the bark covering them was being scratched manually from below to simulate host movement. Some experienced wasps parasitised freeze-killed and EPN-killed hosts when an artificial vibrational stimulus was added in this way and the dead host was moved about inside its chamber. B. hylobii also parasitised hosts the species is not known to be associated with, including larvae of longhorn beetle R. bifasciatum and waxmoth larvae (Galleria mellonella). These results suggest that wasp foraging behaviour, host acceptance and oviposition success is most affected by vibrational cues and that naïve wasps are less likely to locate moribund hosts that move and feed less due to infection with EPN. The risk is expected to be further mediated in a field situation where wasps are likely to encounter several hosts in one host patch and more likely to be attracted to those hosts that are healthy and producing more or stronger vibrational and volatile cues compared to moribund hosts (Chapter VI). In summary, persistence and spread of EPN after application to tree stumps on clearfell sites was low and there was no evidence of a substantial impact on non-target insects. It can therefore be concluded that the overall risks associated with the use of EPN as an inundative biological control agent against the large pine weevil are low