Limited catching bias in a wild population of birds with near-complete census information.

Mirre J P Simons,Shinichi Nakagawa,Isabel Winney,Terry Burke,Julia Schroeder

doi:10.1002/ece3.1623

Abstract

Animal research often relies on catching wild animals; however, individuals may have different trappability, and this can generate bias. We studied bias in mist netting, the main method for catching wild birds. The unusually high resighting rate in our study population—house sparrows (Passer domesticus) on Lundy Island (England)—allowed us to obtain accurate estimates of the population size. This unique situation enabled us to test for catching bias in mist netting using deviations from the expected Poisson distribution. There was no evidence that a fraction of the birds in the population consistently remained uncaught. However, we detected a different bias: More birds than expected were captured only once within a year. This bias probably resulted from a mixture of fieldworkers sometimes ignoring rapid recaptures and birds becoming net shy after their first capture. We had sufficient statistical power with the available data to detect a substantial uncaught fraction. Therefore, our data are probably unbiased toward catching specific individuals from our population. Our analyses demonstrate that intensively monitored natural insular populations, in which population size can be estimated precisely, provide the potential to address important unanswered questions without concerns about a fraction of the population remaining uncaught. Our approach can help researchers to test for catching bias in closely monitored wild populations for which reliable estimates of population size and dispersal are available.

Highlights

Animal research in general, and ornithological research, often relies on catching individuals from a group for monitoring and/or experimental purposes
The chisquare test across all years suggested that the observed pattern of captures deviated from the expected Poisson (v238 = 96, P < 0.001). That these deviations from the expected Poisson distributions do not stem from a subpopulation of birds evading capture in winter
In years 2004 and 2005, there was an overrepresentation of birds that were caught only once or twice, while the Poisson distribution predicted them to be captured more often, resulting in an underrepresentation of birds caught multiple times within a winter

Summary

Introduction

Ornithological research, often relies on catching individuals from a group for monitoring and/or experimental purposes. In captive and wild populations, capture order and phenotypic traits have been found to be positively correlated with, for example, immune functioning (Birkhead et al 1998), sexual signal expression (Birkhead et al 1998; Moreno-Rueda 2003), age (Moreno-Rueda 2003), and growth rate (Biro 2013), suggesting that individuals of higher quality have a lower propensity to be caught Such heterogeneity in catching propensity can severely bias the results and conclusions.

Methods

Results

Conclusion