Does the surrounding matrix influence corridor effectiveness for pollen dispersal in farmland?

Recent decades have seen a shift in agricultural land use from pasture to arable combined with increased use of fertilisers and pesticides. In this quite hostile landscape matrix, pollinator movements between native vegetation remnants may be impeded. Linear landscape elements (LLEs) in farmland can function as biological corridors by facilitating pollinator movements and pollen flow between fragmented plant populations. The type of landscape matrix surrounding LLEs and plant populations, and LLE habitat quality may influence the effectiveness of LLEs as corridors for pollen dispersal through the availability of floral resources and nesting opportunities for pollinators. Using fluorescent dyes as pollen analogues, we investigated dye dispersal patterns between fragmented populations of the rare insect-pollinated Primula vulgaris connected by existing LLEs. We examined how dye deposition on P. vulgaris and within LLEs (on other co-flowering insect-pollinated species) could be influenced by the surrounding matrix type (pasture, arable field), the recipient population traits of P. vulgaris (flowering population size, flower display, flowering plant density and co-flowering floral resources) and by LLE traits (LLE length and co-flowering floral resources). Dye dispersal through corridors was significantly higher when the landscape matrix surrounding P. vulgaris recipient populations and LLEs consisted of pastures rather than arable fields (or a mix of both), even after accounting for differences in floral resources. A higher cover and diversity of insect-pollinated plants increased dye deposition when co-flowering within small P. vulgaris populations, but led to dye loss within LLEs. Large P. vulgaris populations appeared more attractive thereby increasing heterospecific dye deposition in the LLEs. Our study shows that farming practices shifting from pastures to maize arable fields have a negative impact on dye dispersal patterns, and so possibly affect pollen dispersal of P. vulgaris, likely through a reduced pollinator service. Corridor effectiveness depends on the type of surrounding matrix, and the abundance of floral resources directly influences dye dispersal patterns. Preserving the remaining pastureland is essential, but restoring rich flowering vegetation at field boundaries and along LLEs may also promote corridor effectiveness for pollen dispersal.