Multi-level Bayesian analysis of monk parakeet contact calls shows dialects between European cities

Geographic differences in vocalisations provide strong evidence for animal culture, with patterns likely arising from generations of social learning and transmission. The current knowledge on the evolution of vocal variation has predominantly focused on fixed repertoire, territorial song in passerine birds. The study of vocal communication in open-ended learners and in contexts where vocalisations serve other functions is therefore necessary for a more comprehensive understanding of vocal dialect evolution. Parrots are open-ended vocal production learners that use vocalisations for social contact and coordination. Geographic variation in parrot vocalisations 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 multi-level Bayesian model and sensitivity analysis, with this novel approach allowing us to explicitly compare vocalisations 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 vocalisation studies.

compound nests, the latter containing multiple nests, each with one or multiple chambers per pair (SQS personal observation). Nest openings correspond to nest chambers, which can serve as a proxy 122 for population size. These nest structures occur in larger nesting colonies. The term colony is 123 often defined as one or more nest structures located within 200m of each other (see (Reed et al. 124 2014)). In cities and invasive populations, these nesting colonies are often located within parks or 125 other green areas, clearly delineated from other colonies (Eberhard 1998), although with potential 126 between-park movement and dispersal (Bucher, Martin, et al. 1990). A recent study in the native 127 range of monk parakeets found evidence that individual signatures outweighed any emergent dialects 128 (Smith-Vidaurre, Araya-Salas, and Wright 2020). Interestingly, regional dialects between cities have 129 been observed in the invasive populations of monk parakeets in the United States (Buhrman-Deever, 130 Rappaport, and Bradbury 2007). 131 In the current study, we aim to assess these competing hypotheses by examining patterns of   Table 1 for sampling effort per park, see Figure 1 for sampling area, and see supplemental materials  were also videotaped with a Philips HC-V777EG-K to allow assignment of calls during processing. 171 We tested how this incorrect pooling might have affected the results in a sensitivity analysis (see 2021), all selected calls from Raven were clipped and high quality spectrograms were created (see 178 Data availability statement). All spectrograms were then manually inspected and calls that were 179 considered to be poor quality were removed.

180
The remaining calls were categorised as contact calls (tonal calls with at least three peaks in their 181 frequency modulation) and other calls. Contact calls were further manually sorted into six variants:  To assess whether our categorizations of call variants were reproducible, we created a randomized sample of 1000 calls from our dataset, including both contact and non-contact calls. We then asked 192 an independent observer to classify the calls. We assessed both how the observer's classifications   There was clustering by city (see Figure 3b) with distinct differences based on PC1 (see Figure 4c). contrasts (see Figure 3). These results were consistent across methods (see Supplemental Materials).

239
In contrast, there was little evidence for widespread differences between parks within cities.  (see Supplemental Materials, Figure S3). Park level means can appear very different under incorrect 247 pooling, even when no signal exists in the simulated data (see Figure S1, Supplemental Materials).

248
The city level signal we detected is much stronger than the simulated results due to incorrect pooling 249 (see Supplemental Materials, Figure S2). This lends strong support for dialect differences between 250 cities, while there is no support for this at the park level given the few differences observed.

251
In addition to assessing overall differences between parks and cities, we examined the proportion 252 of contact call variants that were observed across the different cities (see Figure 5). We found that in Our results provide strong evidence that monk parakeet contact calls differ between multiple cities 261 across their European range. Vocal differences between the parks within cities were also detected, 262 however, these differences were less consistent compared to the dialect pattern we observed at the 263 city level and appeared to be only present in a few parks (see Figure 4). Overall, our results 264 provide support for the cultural drift hypothesis, while finding no support for the group membership

333
While we did not find evidence for strong convergence towards a group level signature in contact 334 calls, it could be the case that group signatures exist in other call types, or within very specific 335 variants of contact calls. In accordance to our call type analysis, (see Figure 5), most variants were 336 present in all cities, but some showed higher proportions than others. While we cannot be certain 337 that these variants drive the dialect differences between cities, or lack of in parks, they raise an