Artificial light is the main driver of nocturnal feeding by the Rock dove (Columba livia) in urban areas

Artificial light at night (ALAN) is one of the most extreme alterations of urban areas, which drives nocturnal activity by diurnal species. Although the Rock Dove (Columba livia) is a common species in urban centers worldwide known to have nocturnal activity in urban areas, it is unknown what is the role of ALAN in its nocturnal activity. Moreover, studies that address the relationship between ALAN and nocturnal activity of diurnal birds are scarce in the Southern hemisphere. The objectives of this study were: 1) to evaluate the extent of nocturnal activity in the Rock Dove in large cities of Argentina; and 2) to analyze the influence of ALAN, pedestrian traffic and car traffic on the nocturnal activity in two cities, Buenos Aires and Salta. I visited the most urbanized areas of five large cities in Argentina, and surveyed lighted streets once after 30 minutes after sunset. In Buenos Aires and Salta, I compared environmental conditions between sites were doves were seen feeding with random sites were doves were not recorded feeding. Nocturnal feeding of the Rock Dove was recorded in three of five cities surveyed. ALAN was positively related to nocturnal feeding activity in Salta and Buenos Aires. The results obtained suggest that urbanization promotes a nocturnal activity of the Rock Dove, which occurs in cities located in a vast range of altitudes and biogeographic contexts. Moreover, the nocturnal activity is mainly driven by ALAN, which probably alters the circadian rhythm of doves.

street. In Buenos Aires, given its larger size than the other cities, surveys were made three nights on the main 6 6 avenues and the pedestrian street. Surveys started between 33 to 80 minutes after local sunset and lasted 6 7 between 36 to 120 minutes.

8
In cities with the most quantity of nocturnal feeding doves I located a similar number of random sites where 6 9 doves were not seen feeding at night. In both sites light intensity, car and pedestrian traffic were measured 7 0 once. Car and pedestrian traffic were calculated as the number of people and car passing during three minutes.

1
In the case of pedestrian streets, there was no car traffic. Pedestrian traffic was considered an indicator of food 7 2 for doves, because the greater the number of passing people, the greater the chance to obtain food for them 7 3 intentionally or unintentionally. Car traffic was used as an indicator of noise, given that some studies 8 8 8 9

0
Doves feeding at night were observed in three of the five cities surveyed (Table 1). Buenos Aires and Salta 9 1 had the highest number of nocturnal feeding events (eight and four, respectively), whereas in Mar del Plata 9 2 only one nocturnal feeding event was recorded in the pedestrian street of downtown. In Buenos, during each 9 3 5 feeding event between one to 16 doves were involved (mean = 7.5), whereas in Salta between one to three 9 4 doves were seen. In Salta, nocturnal feeding was only observed in pedestrian streets, whereas in Buenos Aires 9 5 was recorded in pedestrian streets and avenues.

6
There was a significant increase in the probability of nocturnal feeding with more light intensity (LRT = 8.27, 9 7 P = 0.004; intercept = -1.84, slope = 0.04). A light intensity of more than 100 lux had the highest probability Results showed that nocturnal feeding of Rock doves occurred in more than half of cities surveyed, which had 1 0 2 a notable biogeographic heterogeneity (from 24 masl to 1187 masl, and a difference of 14 degrees of latitude) 1 0 3 suggesting that could be a generalized behavior among Latin American cities and worldwide judging by 1 0 4 records reported by Luniak (2004) for Europe. However, the present study focused on large cities of more 1 0 5 than 500 000 inhabitants, so the effect of urbanization on nocturnal dove behavior need to be studied in 1 0 6 smaller cities. On the other hand, much more study is needed to elucidate why the nocturnal feeding of the 1 0 7 Rock Dove is more likely in some large cities and absent in others. Possible explanations may be related to 1 0 8 the levels of artificial light at night and the availability of discarded food by humans. Indeed, in three cases of 1 0 9 nocturnal feeding in Buenos Aires two people were actively involved in feeding doves, whereas in one case 1 1 0 doves were feeding on discarded food in a garbage container. Moreover, cities with elevated numbers of 1 1 1 doves may encourage nocturnal feeding as a way to avoid intraspecific competition (Leveau 2018).  Blackbird and the European Robin (Erithacus rubecula), did not find a significant effect of light intensity and 1 2 0 activity before twilight (Ockendon et al. 2009;Clewley et al. 2016). Disagreements among studies may be 1 2 1 6 related to differences in light intensity. For instance, our analysis spanned a mean light intensity between 8 1 2 2 and 257 lux and was conducted in a highly urbanized area composed by high buildings and commercial areas. melatonin secretion was significantly suppressed at minimum light intensities that ranged between 1 lux to 1 2 9 393 lux (Brainard et al. 1982, McIntire et al. 1989, Aoki et al. 1998

3 6
In this analysis, pedestrian traffic was used as an indicator of food availability for doves. However, this 1 3 7 variable did not affect the nocturnal activity of doves. Other variables indicating food availability and not 1 3 8 considered in this study, such as the number of garbage containers or restaurants may be more related to dove 1 3 9 nocturnal feeding. On the other hand, noise was claimed by several studies as determinant of nocturnal 1 4 0 singing in birds (Fuller et al. 2007, Arroyo-Solis et al. 2013). However, this study showed that nocturnal 1 4 1 feeding by doves did not respond to car traffic, a proxy of noise. The results obtained showed that nocturnal activity by Rock Doves is common in Argentina, considering 1 4 5 large cities. The nocturnal activity was recorded in cities located in a vast range of altitudes and latitudes.