Search

Root-feeding insects are now considered to play a greater role in ecosystem proc- esses than previously thought, yet little is known about their specific interactions with host plants compared with above-ground insect herbivores. Methodological difficulties associated with studying these insects in the soil, together with the lack of empirical and theoretical frameworks, have conventionally hindered progress in this area. 2 This paper reviews recent empirical and theoretical developments that have been adopted for studying root-feeding insects, focusing on the non-invasive tech- niques of X-ray tomography and acoustic field detection and how these can be integrated with new mathematical modelling approaches. 3 X-ray tomography has been used for studying the movements of several insects within the soil and has helped to characterize the host plant location behaviour of the clover root weevil, Sitona lepidus . Acoustic detection of soil insects has been used in various managed systems, ranging from nursery containers to citrus groves. 4 Mathematical modelling plays a complementary role for investigating root-feed- ing insects, illustrated by a number of published models. A model is presented for the movement of S. lepidus in the soil, which suggests that these insects un- dergo Levy movements, similar to those recently demonstrated for above-ground organisms. 5 The future directions and challenges for investigating root-feeding insects are discussed in the context of the wider ecosystem, incorporating both above and below-ground organisms.

1. Few entomological studies include soil-dwelling insects in mainstream ecological theory, for example the preference–performance debate. The preference–performance hypothesis predicts that when insect herbivores have offspring with limited capacity to relocate in relation to a host plant, there is a strong selection pressure for the adult to oviposit on plants that will maximise offspring performance. 2. This paper discusses the proposition that insect herbivores that live above ground, but have soil-dwelling offspring, should be included in the preference–performance debate. Twelve relevant studies were reviewed to assess the potential for including soil insects in this framework, before presenting a preliminary case study using the clover root weevil (Sitona lepidus) and its host plant, white clover (Trifolium repens). 3. Maternal S. lepidus preferentially oviposited on T. repens plants that had rhizobial root nodules (which enhance offspring performance) rather than T. repens plants without nodules, despite plants having similar foliar nutritional quality. This suggests that adult behaviour above ground was influenced by below-ground host-plant quality. 4. A conceptual model is presented to describe how information about the suitability for offspring below ground could underpin oviposition behaviour of parental insects living above ground, via plant- and soil-mediated semiochemicals. These interactions between genetically related, but spatially separated, insect herbivores raise important evolutionary questions such as how induced plant responses above ground affect offspring living below ground and vice versa.

1. In contrast to above-ground insects, comparatively little is known about the behaviour of subterranean insects, due largely to the difficulty of studying them in situ. 2. The movement of newly hatched (neonate) clover root weevil (Sitona lepidus L. Coleoptera: Curculinidae) larvae was studied non-invasively using recently developed high resolution X-ray microtomography. 3. The movement and final position of S. lepidus larvae in the soil was reliably established using X-ray microtomography, when compared with larval positions that were determined by destructively sectioning the soil column. 4. Newly hatched S. lepidus larvae were seen to attack the root rhizobial nodules of their host plant, white clover (Trifolium repens L.). Sitona lepidus larvae travelled between 9 and 27 mm in 9 h at a mean speed of 1.8 mm h(-1). 5. Sitona lepidus larvae did not move through the soil in a linear manner, but changed trajectory in both the lateral and vertical planes.

The phenology and some aspects of the biology of the clover weevil Sitona lepidus Gyll. were studied in the field and the laboratory. Oviposition and hatching were temperature dependent. Most eggs were laid in September and October, although there was also an oviposition period in early summer. Hatching was largely confined to late spring and early summer. The relevance of this to control strategies is discussed. Most eggs were laid on the plant but(were very easily dislodged and probably usually hatch on the ground. In no-choice tests, there were as many eggs laid on more cyanogenic as on less cyanogenic plants and cultivars. There was no evidence that the cyanogenic ability or any other cultivar characteristic would be useful in control strategies.

A series of laboratory and field studies were done to evaluate a range of leguminous plant species for their feeding potential by adult weevils of the genus Sitona Germar. (Coleoptera: Curculionidae). Three species of Sitona, S. lineatus L., S. flavescens Marsh. and S. hispidulus F. all of which are found commonly on white clover (Trifolium repens L.) in the UK were offered a range of 11 legume species, T. repens (white clover, cv. Olwen), T. pratense L. (red clover, cv. Marcom), T. fragiferum L. (strawberry clover, cv. Palestine), T. hybridum L. (hybrid clover, cv. Tetra), T. incarnatum L (crimson clover), T. dubium Sibth. (lesser yellow trefoil), Lotus corniculatus L. (birdsfoot-trefoil, cv. Leo), L. uliginosus Schkuhr. (large birdsfoot-trefoil), Melilotus alba Desr. (white melilot), Medicago sativa L. (lucerne, cv. Europe) and M. lupulina L. (black medick) in two laboratory experiments. The weevils were offered a choice of these legumes in one experiment whilst in the other they did not have a choice of food material. These legumes were also sown in the field and a number of measurements of damage, together with counts of Sitona spp., were made. In the laboratory S. lineatus and S. hispidulus favoured some of the legumes to a greater or lesser extent than white clover. S. flavescens was more restricted in its feeding than the other two weevil species. In the field study S. lineatus invaded the experimental area quickly and tended to favour Medicago spp. and Melilotus spp. Later in the year S. flavescens dominated the sitona fauna on the experiment, with the exception of aggregations of S. lineatus on M. sativa and M. alba. In a separate screen of 5 varieties of white clover (cvs Donna, Menna, Kersey, Olwen and Grasslands Huia), cv. Olwen appeared to be the most susceptible to sitona attack.

In plots of perennial ryegrass and white clover, regular treatment with chlorpyrifos and methiocarb more than trebled the clover content of the sward, compared with untreated plots, in just over 2 years from sowing. Chlorpyrifos alone accounted for most of the effect, which was attributed mainly to the control of Sitona lepidus. Combined treatment increased total herbage dry matter yield by 4·24 t ha−1 during this period but had little effect on grass-only plots. Methiocarb appeared to have contributed by assisting earlier establishment of clover, but otherwise had little effect until slugs, which were initially uncommon, became more abundant in the later stages of the experiment. The effect of slug control was most noticeable in a cultivar of low cyanogenic potential. This factor did not influence the effect of chlorpyrifos.

In laboratory experiments to clarify earlier field observations, Sitona lepidus larvae caused significant damage to established clover plants at infestations of approximately 1–2 per plant. When eggs were placed on the soil, cultivars of differing cyanogenic capacity were damaged equally. The plants did not recover when larval feeding ended. Leatherjackets, slugs and ‘lucerne flea’ fed most on the least cyanogenic cultivars, even where no choice was offered. Plants ultimately recovered from damage by leatherjackets. Deroceras reticulatum caused more damage than did Arion fasciatus. The proportion of leaves damaged by lucerne fiea was highest in the least cyanogenic eultivar but also tended to increase in the most cyanogenic cultivar, confirming a field observation. At high population densities lucerne flea was capable of killing some seedlings but was not tested against established plants.

From 1975 to 1979, biotypes of the braconid wasp Microctonus aethiopoides were forwarded from the Mediterranean region to be reared and released in south-eastern Australia for control of the weevil Sitona discoideus. The origins of the various biotypes of the parasitoid are indicated. M. aethiopoides was recovered from most of the liberation sites. At sites where early releases were made, parasitisation rates of up to 62% per parasitoid generation, and initial dispersal of about 10 km per annum were recorded. The conditions influencing establishment and spread of the parasitoid are discussed.