Avian Food Selection With Application to Pesticide Risk Assessment: Are Dead and Desiccated Insects a Desirable Food Source?

Abstract

Past evaluations of pesticide exposure have been conducted with substantial uncertainty regarding avian consumption of contaminated food items. One question is whether birds consume invertebrates that are killed by a chemical application and that may present an increasing chemical concentration as they desiccate. We addressed the research question in two phases. First, a laboratory study was conducted in which wild-caught birds were individually offered three food choices, i.e., live, fresh-dead, and desiccated insect larvae. Second, these same food choices plus live, fresh-dead, and desiccated crickets were presented in study plots in two agricultural crops, i.e., a cornfield and an orchard. The experimental food items were monitored with videography equipment to determine their fate and to compare laboratory and field results. Laboratory results showed that birds have a strong preference for live and fresh-dead prey over desiccated prey, with live prey taken before fresh-dead prey in most trials. The field study revealed a similar preference for live prey over desiccated prey, with preference for fresh-dead prey intermediate to the two other types.

Insecticide residues in Australian plague locusts (Chortoicetes terminifera walker) after ultra-low volume aerial application of the organophosphorus insecticide fenitrothion

“…This is supported by a study addressing avian food preferences for live, freshly dead, or desiccated insects under laboratory and field conditions. That study concluded that, when given a choice, wild birds did not consume desiccated prey but had a strong preference for live insects, followed by freshly dead individuals [12]. In line with this observation, captive white ibis (Threskiornis molucca) have been shown to prefer live locusts over freshly killed or dried locusts [15].…”

“…Scavenging was most likely both by diurnal and, probably more importantly, by nocturnal scavengers/predators, reducing the number of nymphs available for birds. While other studies suggest that predators such as arachnids and ground beetles rapidly remove prey from fields [12], it is likely that small mammals and the significant number of reptile species present in Australia's arid and semiarid regions must also be contributing to scavenging of moribund and dead locusts [1].…”

“…The assessment of risk resulting from pesticide exposure is further complicated by the fact that previous insect collections have not always been designed to mimic the feeding behavior of predatory birds and thus do not accurately reflect their potential dietary intake. Insects are commonly captured by pitfall traps, which may seriously underestimate residue levels by sampling active insects; by sweep netting, which may overestimate the residue level on locusts from contact with contaminated vegetation; or by collecting dead insects that desiccate rapidly, thereby increasing detected residue levels [12].…”

Measured efficacy, bioaccumulation, and leaching of a transfluthrin‐based insecticidal paint: a case study with a nuisance, nonbiting aquatic insect

“…Realistically, avian consumers are likely targeting freshly dead or dying insect prey with movement detection criteria playing a key role. 27 We measured TFL concentrations of 25.3 ± 0.9 ng/caddisfly after 24 h and consider this a more environmentally realistic exposure profile. The European Commission's Biocidal Products Regulation (BPR, Regulation EU 528/2012) provides a NOEC oral,bird (no observed effect concentration) estimate of <0.07 mg kg −1 feed, 72 i.e., 90 caddisflies after a 24 h TFL exposure for a 30 g bird.…”

Opinion of the PPR Panel on the evaluation of methamidophos in ecotoxicology

“…Both the interpretations above are further affected by the fact that application of methamidophos will reduce both the availability and palatability (Stafford et al 2003) of insects for birds. However, the extent of the reduction in insect populations after spraying is variable and the rapidity of its onset is uncertain 13 .…”