No evidence for the Lombard effect: black-collared Starlings sing at higher frequency but not higher amplitude in urban noise.

Abstract

Low frequency traffic noise in cities potentially masks animal signals, interrupting important fitness-related functions including mate choice and territory defense. Birds commonly sing at higher frequencies to avoid such masking. However, birds are also predicted to sing at higher amplitudes in response to increasing background noise. It remains an open question whether observed frequency shifts are an adaptation to avoid signal masking or a passive by-product of singing at higher amplitude. Few studies have tested both amplitude and frequency of wild bird song in the field due to the difficulties of accurately measuring amplitude in the field. We studied the impact of urban noise on songs of Black-collared Starlings (Gracupica nigricollis), a bird which sings from open perches facilitating amplitude measurements. We measured the amplitude and frequency of starling songs across a noise gradient in a dense urban area. We found that starlings increased the minimum frequency but not the amplitude of their songs as the amplitude of background noise levels increased. Contrary to our predictions, we found an inverse relationship between song amplitude and song frequency, with birds singing at lower frequencies as they produced louder vocalizations. This study provides evidence that birds can alter both the volume and frequency of songs independently. Shifts in song frequency may therefore be a direct adaptive response to noise and are not necessarily a simple by-product of increased song amplitude.