Temporal and spatial variation in sex-specific abundance of the avian vampire fly (Philornis downsi)

Full description

Understanding the range and behaviour of an invasive species is critical to identify key habitat areas to focus control efforts. Patterns of range use in parasites can differ temporally, across life stages and between sexes. The invasive avian vampire fly, Philornis downsi, spends the larval stage of its life within bird nests, feeding on developing nestlings and causing high levels of mortality and deformation. However, little is known of the ecology and behaviour of the non-parasitic adult fly life stage. Here, we document sex-specific temporal and spatial patterns of abundance of adult avian vampire flies during a single Darwin’s finch breeding season. We analyse fly trapping data collected across 7 weeks in the highlands (N = 405 flies) and lowlands (N = 12 flies) of Floreana Island (Galápagos). Lowland catches occurred later in the season, which supports the hypothesis that flies may migrate from the food-rich highlands to the food-poor lowlands once host breeding has commenced. Fly abundance was not correlated with host nesting density (oviposition site) but was correlated with distance to the agricultural zone (feeding site). We consistently caught more males closer to the agricultural zone and more females further away from the agricultural zone. These sex differences suggest that males may be defending or lekking at feeding sites in the agricultural zone for mating. This temporal and sex-specific habitat use of the avian vampire fly is relevant for developing targeted control methods and provides insight into the behavioural ecology of this introduced parasite on the Galápagos Archipelago.

Methods

Adult vampire flies were collected using baited McPhail traps hung in trees (Causton et al. 2019; Lincango and Causton 2009). Traps were baited with 150 mL of liquid lure composed of 600-g ripe Hawaiian papaya, 75-g sugar, and 4 L of water, blended and fermented in the sun 3 days prior to use. Trapping occurred in the highland and lowland site. At each site, traps were placed within four study plots, each containing 12 traps separated by 50 m in a three by four trap lattice (Fig. 1). In addition, four traps were placed in two more study plots along a single transect each separated by 50 m (N = 32 traps per site, total N = 62 traps; Fig. 1). Traps were placed alternatively at 4 and 7 m high to capture potential sex ratio differences of flight height found previously by Kleindorfer et al. (2016). Bait lure was replaced and all specimens collected every 5 days. This was repeated nine times from January 19th to March 5th for a total of 563 trapping events. Collected flies were stored in 70% ethanol, identified, and sexed under a stereomicroscope following morphology described in Kleindorfer et al. (2016). Distance of each trap to the agricultural zone boundary was calculated using coordinate data. The host nesting density, i.e. the number of active Darwin’s finch nests per 200 m × 100 m study plot, was collected from our long-term nest monitoring protocol (see Kleindorfer et al. 2014), which occurred concurrently with trapping. Search effort for active nests within study plots was equal across highland and lowland sites. The host species monitored were the small ground finch (Geospiza fuliginosa), cactus finch (Geospiza scandens), small tree finch (Camarhynchus parvulus), medium tree finch (C. pauper), and the hybrid tree finch (C. parvulus × C. pauper as well as introgressed individuals). Each monitored host nest that was with eggs (incubation phase) or nestlings (feeding phase) within each study plot (100 m × 200 m) during each trapping period (5 days) was counted as an active nest, giving a nesting density of Darwin’s finch nests per plot for each trapping event.