Experimental simulation of some visual and physical components of a hedge and the effects on butterfly behaviour in an agricultural landscape

Experiments on the corridor function of linear features in real landscapes are difficult and expensive to carry out using the classic replicated and controlled experimental regimes usually required as formal proof of a concept. We describe two experiments using model structures to simulate different attributes of hedges for their potential role in initiating corridor movement. We examined: (i) the visual signal or link provided by flowering plants associated with hedges (represented by a narrow red-and-white builder's warning tape) and (ii) the physical presence of the body of the hedge itself (represented by green horticultural windbreak). Removable, 100 m long, links were created between two habitat patches (Angstrom kerholmer) in an arable field in Vestby, Norway. We recorded the effect of the links on the behaviour of three 'species' of the free-flying butterfly community of the study area: the scarce copper (Heodes virgaureae), the heath fritillary (Mellicta athalia) and, as a complex, the high brown (Fabriciana adippe) and niobe fritillaries (F. niobe). The visual link did not elicit a uniform response by the different 'species'. Significantly more high brown/niobe fritillaries flew along the visual link than along the control (no link) and mean distances traveled along the tape were significantly longer than for the control. The heath fritillary and scarce copper did not react in this way. The tape had a slight, but statistically significant, barrier effect on the scarce copper but did not act as a barrier to the other species. All three 'species' were strongly affected by the presence of the physical link. A significant barrier effect was detected for all three 'species' as was a corridor effect with an increased proportion of butterflies moving along the windbreak compared with the control. Mean distances traveled along the windbreak were significantly longer compared with the control, the response being strongest for the scarce copper. The results demonstrated that simple structures could modify the behaviour of species moving through the landscape and that individual species reacted in different ways to the same structure. The experimental approach of using model structures may have promise for modeling aspects of landscape elements in natural situations for a range of species. The results are discussed in relation to butterfly population structure, mate finding, dispersal and corridor design.