Field boundaries play an important role as refuges, food sources and corridors for invertebrates and vertebrates, and increasing farmland fragmentation impacts on these functions. However, hedgerows and other structures can also impede dispersal by flying insects. The current work uses the pollen of Phacelia tanacetifolia in hover fly guts as a marker to assess hover fly movement in farm landscapes. In the United Kingdom and New Zealand, Phacelia pollen was found in the guts of Ephisyrphus balteatus and Metasyrphus corollae (United Kingdom) and Melanostoma fasciatum (New Zealand) at distances up to 200 m from the source, when there were no barriers between the flowers and the traps used to catch the flies. The rate of decline over distance in the proportion of flies containing pollen was similar for the two countries. The extent to which four replicated field boundary types impeded hover fly movement was tested using post-and-wire fences, lines of poplars (Populus spp.) with gaps, dense poplars and controls (no potential barriers). Phacelia was planted on one side of each boundary, and along the centre of the control plots. The relative presence of the pollen in flies on both sides of the barriers showed that both types of poplar boundary restricted the movement of the flies, but the fence had no effect. In a separate experiment, gravid females of M. fasciatum were captured at a greater height on a shade-cloth fence than were non-gravid females and males. The implications of this work include the functioning and persistence of metapopulations and the influence of field boundaries on population recovery of beneficial invertebrates following pesticide-induced mortality. If field boundaries contribute to a temporal asynchrony between pest and natural enemy populations, this needs to be considered along with the well-established roles of boundaries as refugia for, and sources of, beneficial arthropods.

No abstract

Pollen feeding by adult hoverflies of Melanostoma fasciatum, collected from organic paddocks at Flock House, Bulls, North Island, New Zealand, were studied by dissection. Up to six different pollen types were found in one hoverfly. On average, they fed only on 2.71 pollen species. 71% of gravid females but only 13% of males reached the maximum score (scale 0-5) for pollen content in the gut. Very few hoverflies had no pollen. Of the 15 identified pollen groups, plantain, phacelia, coriander, and graminaceous pollen were consumed in large quantities by gravid females. Differences in feeding spectra between sexes probably reflected the need for energy in males and for the protein required for egg maturation in females. Earlier views of strict preference for anemophilous pollen by M. fasciatum cannot be upheld.

Summary1 An experiment was conducted in a winter wheat field using yellow water traps at crop height and at ground level, near to and distant from flowers, to test the hypothesis that such traps are seen as a source of food by flower‐feeding adult hoverflies and are therefore likely to selectively trap hungry individuals. Hoverflies caught in each trap were counted and identified and the amount of pollen in their guts was assessed. Ratios of numbers of hoverflies seen in the wheat crop to numbers caught in nearby traps were compared for the different treatments.2 Most hoverflies were caught in crop‐high traps but they included a high proportion of individuals with empty guts. The taxa were: Episyrphus balteatus (De Geer), Metasyrphus corollae (F.) (these species together accounted for over 90% of the individuals trapped), with Sphaerophoria spp., Syrphus spp., Scaeva pyrastri (L.), Melanostomini and unidentified others making up the rest.3 Significantly fewer hoverflies were captured in low than in high traps. The ratios of numbers trapped to numbers observed, in flower and in no flower treatments would be expected to be the same if the traps were not selective. This was supported for low traps. With high traps, however, there was a highly significant difference between the ratios (71.34 and 126 : 8, respectively).4 Flies captured in high traps had less pollen in their gut than those captured in low ones. At each distance, more E. balteatus captured in high traps were in pollen category 1 (< 20 grains) than in any other category. The opposite state was seen in low trap catches, where most flies were in category 5 (> 5000 grains). Median pollen categories were 2 (21–200 grains) and 4 (1501–5000 grains) for flies caught in high and low traps, respectively.5 The ecological selectivity of traps according to their height and the physiological condition of the targeted individuals is a problem likely to affect many trapping systems apart from the one described in this paper.

Dispersal within agricultural fields and the effects of different barriers on between‐field movement of the New Zealand hover fly Melanostoma fasciatum were studied using ingested pollen as markers. Hover flies did not generally disperse more than 20 m from the pollen source. Gravid females had no significant wind‐directed movement pattern whereas males significantly flew downwind. Flies tended to avoid flying over barren land: a dirt track, an asphalt road or a ploughed field all seemed to hamper hover fly dispersal equally. The implications for spatial arrangement of the flowering strips to enhance the biocontrol potential of hover flies are discussed.