Eumelanic coloration and age interact to influence breath rate following a boldness test in urban pigeons

The rapid urbanization of the past decades has forced numerous species to adapt to their new environments over a very restricted time scale. Previous studies suggested that individuals living in urban areas have specifi c characteristics as compared to those living in rural areas. In feral pigeon populations (Columba livia), individuals living in cities are more melanic than those living in less urbanized areas. In this study, we tested whether the degree of eumelanin-based coloration in pigeons refl ects certain adaptations to urban areas such as boldness and tolerance to stress. Hence, we examined the relationship between eumelanin-based coloration and three variables of boldness (arrival to a new food source, arrival latency, and fl ight distance) and tolerance to stress (breath rate). Our results show that the eumelanin-based coloration of individuals is not correlated with the boldness variables, but tends to be positively related to breath rate in adults. Therefore, eumelanin-based coloration could refl ect the capacity of adult individuals to manage stress in an urban environment, but not the boldness. The higher frequency of eumelanic pigeons observed in cities might thus be due to urban selective forces that favor individuals with a higher response to stress than bolder individuals. Research Article Eumelanic coloration and age interact to infl uence breath rate following a boldness test in urban pigeons Sophie Dupont1, Emmanuelle Baudry2, Pauline Juette2 and


Introduction
Nowadays, urban areas constitute environments of particular interest in ecology and evolution. The rapid urbanization of the past decades as part of global change has forced numerous species to adapt to these new environments over a very restricted time scale. In the context of growing worldwide urbanization (increase of up to 1.2 million km 2 between 2000 and 2030, [1], it becomes crucial to understand how natural species adapt to this new environment and, more generally, how this ecosystem functions. While few studies focus on the adaptation of populations to the urban environment, they suggest that individuals living in urban areas have specifi c characteristics as compared to those living in rural areas [2]. The combination of particular traits would enable individuals to successfully exploit urban environments. This urbanization syndrome or synurbization [3] may include high reproduction rates, higher levels of aggressiveness and boldness, attenuated physiological responses to stressors, and increased tameness toward humans [5]. Moreover, the degree of melanin-based coloration has been frequently reported to covary with life-history, morphological, physiological, and behavioral traits, as well as with the sign and magnitude of these covariations fl uctuating in time and space [6]. These results suggest that the degree of melanin-based coloration might refl ect adaptations to different environmental conditions [7]. A recent review highlighted potential candidate genes pleiotropically affecting both the degree of individual melanization and other phenotypic traits, including aggressiveness, tolerance to stress, energy homeostasis, and anti-infl ammatory immune reactions [8], suggesting that melanin-based coloration could be associated with several aspects of the urbanization syndrome. Accordingly, a recent study in urban pigeons (Columba livia) showed that the degree of eumelanin-based coloration was positively correlated with the urbanization rate [9]. If this hypothesis is true, we should observe a correlation between eumelanin-based coloration and certain aspects of the urbanization syndrome.
In this study, using the feral pigeon (Columba livia), we tested this prediction by investigating the association between the eumelanin-based coloration of pigeons and two potential aspects of the urbanization syndrome: boldnessthe tendency of individuals to take risks, particularly in novel contexts -and tolerance to stress. Urban pigeons display alternative melanin-based coloration, which varies in terms of both melanin types and pigmentation patterns [10]: red coloration is due to pheomelanic pigments and black coloration to eumelanic pigments. These two types of melanic pigments can vary in intensity in feathers, ranging from a total absence of pigmentation (white coloration) to intermediate and even dark melanin coloration. In Paris (France), a preliminary observational study based on photos (n = 3041) revealed that pheomelanic morphs are rare (7 %) as compared to eumelanic morphs (93 %) (Lisa Jacquin and Julien Gasparini, personal observation). The main color polymorphism concerns the quantitative variation of eumelanin present in feathers [9]. A cross-fostering experiment also showed a strong heritability (0.88), suggesting a strong genetic control of this phenotype [11].
We used a boldness test in captivity to quantify the behavioral variables of pigeons (i.e., arrival to a new food source, latency of this arrival, and fl ight distance) as well as a physiological variable related to stress tolerance (i.e., breath rate) and then correlated these variables with eumelanin-based coloration, the main color polymorphism in pigeons.

Material and methods
In and peas, as well as mineral supplements and water. Upon their arrival at the aviaries, the birds were weighed and their coloration score determined (percentage of black coloration on wings) following the method described in Jacquin, et al. [12].
This score is a good approximation of the eumelanin content of feathers [13].
The subsequent testing period lasted 8 days: 3 days of environmental acclimation and 5 days of testing. During this period, all individuals were housed separately in a new aviary and fed with a restricted diet. Individuals were given a daily diet of 10 grams of a wheat and corn mix (delivered every morning).
Because of the restricted caloric intake of this diet and the stress associated with captivity, individuals' weight was regularly monitored. Individuals whose weight decreased too much were excluded from the experiment and re-fed ad libitum. During the 5 days of testing, birds were submitted to two boldness tests followed by two breath rate measurements, the fi rst in the morning and the second in the afternoon (78 trials for 39 individuals). For the boldness tests, two protagonists were considered, namely a feeder who was known to the pigeons and an unknown observer. First, the feeder placed some wheat and corn mix on the fl oor of the aviary. The latency (in seconds) between the setting down of the food and the beginning of its consumption was measured using a chronometer. The trial was stopped if the animal did not eat the food after a time lapse of 20 minutes. When the individual started to eat, the observer, initially located 10 meters away from the aviary, walked in straight line toward it. As soon as the pigeon moved away from the seeds, the observer stopped walking. The distance between the observer and the food source was called the fl ight distance. After measuring the fl ight distance, each individual was captured to evaluate its stress level based on the breath rate [14]. The time taken to capture the bird (latency of capture) was noted because of its possible effect on this measurement. The number of respiratory movements in 1 minute was determined by placing the pigeon on its back in the hands of the observer. In summary, the 78 trials allowed us to obtain 4 variables of interest for 39 individuals: the arrival (or not) of the birds at the food site, the latency of this arrival, the fl ight distance, and the breath rate (a proxy of stress during this test).
We performed four mixed models to investigate whether the coloration and age (adults/juveniles) of birds and their interaction impacted the four variables of interest. All statistical analyses were performed using SAS (version 9.4). First, the arrival was coded as a binary variable (0: the bird did not come to eat the food; 1: the bird came to eat food). We used a mixed model with binomial distribution (logit link function) with the arrival as the dependent variable and the age (juveniles vs. adults), the period of the day where the trial has been done (morning vs. afternoon) and eumelanic coloration (percentage of black coloration on wings) in addition to their interaction as the explanatory variables. We included the location of capture and the individual nested within the location as random factors to take pseudoreplication into account. Then, similar mixed models were run on the logarithm of the arrival latency, fl ight distance, and breath rate with normal distributions. We log-transformed the arrival latency to better satisfy the assumptions of normality. For the breath rate model, we also included the latency of capture as a covariate. Non-signifi cant interactions were removed step-by-step to obtain the fi nal model.

Results
Among the 78 trials (n = 39), the birds decided to go to the food on 68 occasions. This decision was not affected by the age or coloration of the pigeons (Table 1), and only one individual never went to food. However, pigeons decided to go to the food more often in the afternoon (38 times over 39) as compared to the morning (30 over 39; Table 1). We then investigated the arrival latency, fl ight distance, and breath rate for these 68 trials involving 38 birds. We fi rst found a positive relationship between arrival latency and fl ight distance (F 1,28 = 6.05, P = 0.02, Figure 1). However, these two variables were not correlated with the breath rate (mixed models; arrival latency: F 1,28 = 1.33, P = 0.26; fl ight distance: F 1,29 = 0.12, P = 0.73). The arrival latency was not affected by coloration and age of the birds whether individually or in interaction (Table 1) but tend to be longer during trials done in the morning (365 seconds ± 83) than in the afternoon (134 seconds ± 42; non-signifi cant trend, Table 1). The fl ight distance was longer for adults (206 cm ± 21, n = 24) than for juveniles (149 cm ± 25, n = 14; Table 1) and was longer during trials done in the morning (303 cm ± 39) than during trials done in the afternoon (185 cm ± 16; Table 1).
The breath rate was signifi cantly affected by the arrival latency and by the interaction between age and coloration ( Table 1).
The longer the time necessary to catch birds, the higher their breath rate was. In adults, darker pigeons tended to have higher breath rates than paler ones (F 1,20 = 3.66, P = 0.07; Figure 2a).

Discussion
In the present study, we wanted to test whether a darker eumelanin-based coloration in pigeons refl ects their better adaptation to urbanization, namely in terms of greater boldness and tolerance to stress. Our experiment provided contrasting results. When considering boldness variables and contrary to previous studies in other species [15], the results did not support our hypothesis as eumelanin-based coloration was not related to arrival, arrival latency, or fl ight distance (Table 1). In contrast, the breath rate, as a proxy of tolerance to stress, tended to be higher in darker adults, a trend that we did not observe for juveniles. This suggests that eumelanic adult pigeons tend to have a higher response to stress than paler ones immediately after capture and therefore a lower capacity to cope with stress. In contrast with our results, breath rate was found to be negatively correlated with the width of eumelanic black tail band in kestrels (Falco tinnunculus) [16] and with the eumelanic black spots numbers in barn owls (Tyto alba) [17,18].
Overall, our results are rather contrasted with previous studies [8,19] but it is possible that the direction of the relationship between response to stress and melanin-based coloration might depend on species as observed for others physiological component [8].
In our species, darker eumelanic individuals are more prolifi c in highly urbanized areas [9], suggesting that they are potentially better adapted to urban areas than paler ones. From this result, we hypothesized that one of this adaptation would be a better capacity to cope with stress for darker eumelanic pigeons and we predicted that darker eumelanic pigeons would have a better capacity to cope with stress as compared to paler ones. We tested this hypothesis in the present study, but our results found a relation opposite to our prediction. Indeed, in our study, paler pigeons tend to better tolerate stress following capture, at least in adults. Our results are nevertheless consistent with a recent study in pigeons showing that corticosterone response, as a physiological measure of stress, increased with the degree of eumelanin-based coloration in a rural population [20]. Previous studies suggest that urbanization selects animals that are more tolerant to stress [21]. Yet perhaps this speculation is incorrect, as urbanization would appear to select individuals that maintain a high level of stress response. Indeed, a higher stress response may have a positive impact on fi tness in certain habitats such as those with a high predation rate [22]. Urban areas may also include several environmental stressors (anthropic stressors) that select individuals with a high stress response. In our pigeon population, such individuals seem to be eumelanic, which could explain their higher frequency in highly urbanized areas.
This interpretation is based on our correlative result ( Figure   2a). Future studies should thus experimentally validate this scenario and identify the environmental stressor in urban areas that favors individuals with a higher stress response. Even if the boldness measurements (arrival, arrival latency, and fl ight distance) were not correlated with eumelanin-based coloration in pigeons (our main hypothesis), we found an interesting positive relationship between arrival latency and fl ight distance (Figure 1). This result confi rms that these two variables measured the same behavioral syndrome, notably boldness. Flight distance is a common method used to estimate boldness in birds and is applied in many studies [23,24]. By contrast, arrival latency is used less often when investigating Table 1: Output of the generalized mixed models explaining variations in the arrival (arrived or not, binary distribution), the logarithm of the arrival latency, the fl ight distance, and the breath rate (normal distribution) of birds according to their ages (juvenile or adult), the time of the day (trials done in the morning or in the afternoon), eumelanic coloration, and interaction. For the breath rate, capture latency was added to the model as a covariate and was therefore not avalaible (NA) for the three other variables. Nonsignifi cant interactions were removed step-by-step (dash sign) to obtain the fi nal model.  [25,26]. Hence, the food restriction of our study might explain why most of birds took the risk of going to the food. Accordingly, pigeons were more prone to go to the food in the afternoon than in the morning. Indeed, birds were daily fed in the early morning and therefore they were more starved in the afternoon than in the morning. Globally, we observed that this higher starvation increases the risk-taking:

Arrived or not
a higher frequency to go the food, a shorter latency to go the food and a shorter fl ying distance (

Conclusion
Our study shows that the eumelanin-based coloration of individuals could refl ect the capacity of adults to manage stress in an urban environment, but not boldness. Therefore, the higher frequency of eumelanic pigeons observed in cities might be due to urban selective forces favoring individuals with a higher response to stress as opposed to bold individuals.
Boldness is therefore not an aspect of the urbanization syndrome. By contrast, a higher response to stress could be part of the urbanization syndrome in pigeons.

Ethical note
This study was carried out in strict accordance with

Data sharing and data accessibility
The data will be deposited in Dryad or other equivalent archives when accepted.