Food-choice experiment
To assess food preferences of individuals, we presented each pigeon with a six-section circular Plexiglas feeder containing 10g of six different seed types (green peas, oats, popcorn maize, soybeans, sunflower seeds, and wheat), an amount large enough to avoid being depleted during the trial. The six food types were part of the diet of pigeons during the acclimation period, and they were chosen to provide variation in size and nutritional contents. Individuals were allowed to feed for 20 minutes. After this time, we removed the feeder and measured the amount of each food remaining with a digital precision balance. The amount of each seed type consumed per individual was estimated by subtracting this quantity from the initial 10g. Trials in which an individual did not eat were excluded from the analyses. To evaluate the consistency in the measure, each individual was tested in a battery of 4 trials conducted in two consecutive days. In the first daily trial, individuals were tested under food-deprived conditions whereas in the second they had already eaten. Thus, our experimental design allowed us to assess the extent to which fasting influenced food choice (Moon & Zeigler 1979). For some individuals (12 from MoiĆ , 14 from BCN and 28 raised in captivity), long-term consistency in food preferences was evaluated by repeating the food choice experiment a year later (two trials in one day).