Multilevel Bayesian analysis of monk parakeet contact calls shows dialects between European cities

Abstract

Geographic differences in vocalizations provide strong evidence for animal culture, with patterns likely arising from generations of social learning and transmission. Most studies on the evolution of avian vocal variation have predominantly focused on fixed repertoire, territorial song in passerine birds. The study of vocal communication in open-ended learners and in contexts where vocalizations serve other functions is therefore necessary for a more comprehensive understanding of vocal dialect evolution. Parrots are open-ended vocal production learners that use vocalizations for social contact and coordination. Geographic variation in parrot vocalizations typically take the form of either distinct regional variations known as dialects or graded variation based on geographic distance known as clinal variation. In this study, we recorded monk parakeets (Myiopsitta monachus) across multiple spatial scales (i.e., parks and cities) in their European invasive range. We then compared calls using a multilevel Bayesian model and sensitivity analysis, with this novel approach allowing us to explicitly compare vocalizations at multiple spatial scales. We found support for founder effects and/or cultural drift at the city level, consistent with passive cultural processes leading to large-scale dialect differences. We did not find a strong signal for dialect or clinal differences between parks within cities, suggesting that birds did not actively converge on a group level signal, as expected under the group membership hypothesis. We demonstrate the robustness of our findings and offer an explanation that unifies the results of prior monk parakeet vocalization studies.

Keywords: Bayesian statistics; communication; culture; dialects; monk parakeet; open-ended vocal learning.

Figure 1

Map of study locations. Map created using ggmap (Kahle and Wickham 2013), ggrepel (Slowikowski 2021), and ggsn (Santos Baquero 2019). Map tiles by Stamen Design, under CC BY 3.0. Data by OpenStreetMap, under ODbL.

Figure 2

Examples of four contact call variants. a) typical, b) four triangle, c) ladder start, d) ladder middle, e) ladder multiple, and f) mix alarm.

Figure 3

Result for PCO. Colors represent cities (see legend). a) City (thick) and park (thin) averages (dots) and 50, 90, and 95% intervals for PC 2. b) Scatter-plot of all calls included in the model. c) City (thick) and park (thin) averages (dots) and 50, 90 and 95% intervals for PC 1.

Figure 4

Pairwise contrasts between city means in principal coordinate space. Numbers in brackets give the 89% posterior interval for principal coordinate 1 (top) and 2 (bottom) for all city pairs. Intervals are in bold if they do not overlap 0. Squares are colored dark if either one (light blue) or both (dark blue) of the intervals do not overlap 0.

Figure 5

Distribution of variants across cities. Numbers are represented in black, colors represent the percentage of the given variant within the given city and range from 0% (blue) to 100% (orange)—see color scale bar.