Finding the best angle: pigeons (Columba livia) weight angular information more heavily than relative wall length in an open-field geometry task

Abstract

Pigeons were trained to locate food in two geometrically equivalent corners of a parallelogram-shaped enclosure. Both the angular amplitude of the corners and the length of the walls alone were sufficient for successfully completing the task. Following training, birds were tested in three separate conditions that manipulated the geometric information available. During tests in both a rectangular-shaped enclosure that preserved the wall length information but not the angular amplitude, and a rhombus-shaped enclosure that did the opposite, pigeons located their goal corners with a high degree of accuracy, indicating an ability to use both types of geometric information in isolation. This result is consistent with prior research with domestic chicks. However, in a conflict test in a reverse parallelogram-shaped enclosure, in which the correct angular location was paired with an incorrect wall length location, birds showed a preference for the correct angular location. This suggests that pigeons weight angles more heavily than wall lengths in this type of navigation task, which differs from findings in a similar task conducted with the domestic chick. Results in the conflict test also suggest that pigeons did not use the principal axis as their main strategy of small-scale navigation.

Geometric orientation by humans: angles weigh in

“…Although one might argue that participants used the principal axis rather than angles in the rhombus environment, the salience of the principal axis was drastically reduced in the rhombus as compared to the rectangle, and therefore the use of a principal axis strategy should have led to a difference in accuracy between these tests. However, the present results showing that humans performed equally well in the rhombus and rectangle test environments are consistent with those found in similar tests with both pigeons (Lubyk & Spetch, 2011) and domestic chicks (Tommasi & Polli, 2004). This finding is interesting and suggests considerable cross-species generality.…”

“…Moreover, preference for the correct angular information was similar, whether participants were trained to find an acute corner (60°) or an obtuse corner (120°). These results are also consistent with those reported for pigeons (Lubyk & Spetch, 2011), but not for chicks (Tommasi & Polli, 2004); chicks trained to locate acute angles chose the correct angles, whereas chicks trained to locate obtuse angles chose the correct relative wall lengths. Tommasi and Polli suggested that acute angles may be more visually salient than obtuse angles, and therefore, in a conflict situation, chicks would rate the acute angle higher than relative wall length but would rate the obtuse angle lower than the wall length information.…”

“…In the conflict test, this would bring obtusetrained participants to choose the acute corner and acute-trained participants to choose the obtuse corner, which is the opposite of our results. Recent results with pigeons (Kelly, Chiandetti, & Vallortigara, 2011;Lubyk & Spetch, 2011) and domestic chicks have also been inconsistent with a principal axis strategy. Newcombe et al (2010) also found that two-year-old children were able to orient in an octagon-shaped room that contained only obtuse angles and had no single principal axis of symmetry.…”

“…However, when these geometric properties were placed in conflict, chicks trained to choose the smaller (i.e., acute) corners weighted angles more heavily than wall lengths, whereas chicks trained to locate the larger (i.e., obtuse) corners weighted the wall lengths more heavily. In a similar task, Lubyk and Spetch (2011) found no effect of training angle in pigeons: Pigeons weighted angles more heavily than relative wall lengths whether they had been trained to locate acute or obtuse corners. Finally, Jones, Pearce, Davies, Good, and McGregor (2007) demonstrated that rats can use angular amplitudes or wall length relationships to orient in a kite-shaped arena.…”

Reorienting in Virtual 3D Environments: Do Adult Humans Use Principal Axes, Medial Axes or Local Geometry?

“…A growing number of studies have shown that reorientation by human and non-human animals is highly sensitive to alterations in local geometry and our results add to this body of research [7], [10], [11], [28], [29]. During testing in the L-shaped environment, participants directed the majority of their searches to the corners that matched the medial axes of the training environment.…”