European Starling Nest‐site Selection Given Enhanced Direct Nest Predation Risk

Abstract

The European starling (Sturnus vulgaris) is recognized, outside its native range, as an invasive species, and efforts to deter its nesting have generally been unsuccessful. Enhancing perceived risk at the nest site is a little‐explored route in developing a nest deterrent. Specifically, nest predation risk affects site selection, reproduction, antipredator behavior, and fitness in nesting birds. We questioned whether perceived predation risk as related to a snake model positioned inside a nest box, in striking form, moving, and present throughout the period of cavity site selection would induce European starlings to avoid or delay nest construction. We conducted our study on the 2,200‐ha National Aeronautics and Space Administration Plum Brook Station (PBS), Erie County, Ohio, USA. Our objectives were to examine nest initiation, clutch development, and hatching success by starlings relative to 4 nest‐box treatments: moving snake model, movement mechanism only, immobile snake model, and control (empty box). We applied treatments across 120 nest boxes on PBS (30 boxes/treatment) from 15 April through 17 June 2019. We observed no treatment effects on nest initiation or date of nest bowl formation. Starlings occupied (laid ≥1 egg) in 76% of nest boxes (moving snake model = 22 boxes, movement mechanism = 25 boxes, immobile snake model = 21 boxes, control = 23 boxes). We observed no effects of treatment on likelihood of starlings laying a clutch, date of first egg, clutch size, or hatching success. We suggest that starlings adapted to perceived risk because models, despite animacy, did not attack. Therefore, snake models, as developed herein, are ineffective starling nesting deterrents. Published 2021. This article is a U.S. Government work and is in the public domain in the USA.

Integration of multimodal cues does not alter mean but reduces variance in avian responses to predators: a systematic review and meta-analysis

“…Iglesias, et al 30 No control Jones and Sieving 31 Species level data cannot be extracted for control/treatment contrasts Journey, et al 32 No s.e. (standard error) provided for species level data Keen, et al 33 Social learning of novel cue Kerman, et al 34 No control Leavesley and Magrath No treatment: study tests whether trills convey predator information Macleod, et al 36 Response variable is cumulative mass gain expressed as percentage (i.e., control and treatment scaled to same range from 0 to 100) Madden, et al 37 mobbing calls used for treatment were generated in response to humans Martinez, et al 38 Could not extract species level data-figure resolution too low to extract overlapping data points McIntyre, et al 39 Effect direction not extractable Morosinotto, et al 40 response variable = testosterone excluded because no clear predicted effect (authors themselves stated no single prediction) Nilsson and Nord 41 Not relevant -no predator treatment Nocera and Ratcliffe 42 Mixed species level data Pascual and Senar 43 Manipulation is distance to cover Poysa, et al 44 Not relevant -no predator treatment Rajala, et al 45 No control Rands and Cuthill 46 manipulation is human threat Roncalli, et al 47 Response variables not relevant (egg touches, egg rejection) Schneider and Griesser Cannot extract behavioural response to treatments in isolation Serra and Fernandez 49 manipulation is human threat Sieving, et al 50 Response variables is structure of acoustic response -species specific, not generalizable Thompson, et al 51 Not relevant -no predator treatment Tilgar and Moks 52 Mixed species level data Tilgar, et al 53 manipulation is human threat Tolvanen, et al 54 Not relevant -no predator treatment Turney and Godin 55 Mixed species level data Tvardikova and Fuchs Mixed species level data Williams and Lindell 57 Not relevant -no predator treatment Williamson and Fagan 58 Mixed species level data Zanette, et al 59 Response subcategory (cFOS levels) not used in any other study, also, no clear directional prediction…”