Preliminary investigation of the representation of visual proportions in 4-day-old domestic chicks (Gallus gallus).

Animals can solve rudimentary mathematical operations with discrete quantities (Vallortigara et al., 2010). Studies involving discrimination of continuous quantities are scanty (e.g., vanMarle et al. 2006; Beran, 2010) and seem to suggest that representation of continuous quantities may be performed similarly to representation of discrete quantities. Here we wanted to investigate if day-old domestic chicks (an animal that are known to master numerical discriminations, ordinal comprehension and arithmetic calculations; Rugani et al., 2010, 2011, 2013) can represent proportion of continuous quantities. During training, three-day-old chicks were presented with two stimuli, each characterized by different and complementary proportions of red/green areas (\ubc vs. \ube in an overall area of 16 cm2). Food could be found in proximity of only one of the two stimuli: Half chicks were trained to respond to the \ube green area and the other half to the \ube red area. At test, on day 4, both groups of chicks (\ube red: n=10; mean=79,000, SE=3,712; p<0.001; \ube green: n=10; mean=69,000, SEM=4,469; p<0,001) approached the proportion associated with food, even if at test the spatial disposition of the two areas was novel with respect to what experienced at training. In this task chicks could respond to an absolute area and not to a proportion. In Experiment 2, chicks (n=10) trained with the same stimuli used in Experiment 1, were required to generalize to test stimuli of enlarged dimension, that created a conflict between the absolute positive area experienced during training and the correct relative proportion between the two areas. Chicks responded on the basis of the proportion, ignoring the absolute values (mean=78,000, SE=2,486; p<0,001). Even when, in Experiment 3, chicks (n=10) at test were required to respond to new couple of stimuli, each depicting different numbers of little squares on a white background (i.e. 1 green & 3 red vs. 3 green & 1 red or 5 green & 15 red vs. 5 red & 15 green), they were again able to respond to the correct proportion (mean=82,000, SE=2,494; p