Aspects of hens’ visual behaviour

Five experiments were conducted to examine different aspects of hens’ visual behaviour, and to assess whether hens respond to slide images in the same way as they respond to real conspecifics that those slides depict. Experiment 1 was a replication of a study by Candland (1969). Six domestic hens were trained to make simultaneous discriminations between slide images of conspecifics that differed only in combs, beaks and wattles. In further test trials these facial appendices were rearranged in all the possible combinations to determine which was the most salient. The comb appeared to be the most salient feature in both Candlan’s study and Experiment 1. In Experiment 2 the hens were retrained to discriminate between images of two different conspecifics, and then presented with the real object hens that the slides depicted. The subject hens did not discriminate between the real conspecifics. During Experiment 3 the same subject hens were presented with a series of different images, under the same conditions as Experiment 1. Test trials revealed that most hens discriminated between pairs of images on the basis of the uppermost feature, irrespective of whether that feature depicted part of a conspecific or a geometric shape. Backprojected sinusoidal grating patterns of varying contrast and spatial frequency were used to assess the visual resolution of four hens in Experiment 4. Contrast sensitivity functions were measured under two different viewing conditions. These two functions both spanned a range of low spatial frequencies, that is, showed that the hens had low visual acuity, and also differed in the ranges of spatial frequencies, indicating different levels of visual acuity for frontal and lateral viewing. Experiment 5 was designed to assess and demonstrate high-frequency image loss in hens. Four hens were trained to make simultaneous discriminations between two copies of one of the images used in Experiment 1. One of the copies had been blurred by removing high-frequency contrasts. During a subsequent test phase, the hens were presented with pairs of this image, with one of each pair blurred by removal of high-frequency contrasts across a range of cutoff frequencies. A mean threshold for discrimination of high-frequency image loss was measured at 2.237 cyc/° (cycles per degree of visual angle). Demonstrations of this image loss are shown in Experiment 5. Low visual resolution was measured for the hens - at around the same levels as pigeons and six-month old humans. It was concluded that hens do not respond to slide images in the same way as they respond to real conspecifics that those slides depict, but that they might learn to do so after extensive discrimination training with both types of stimuli. Possible influences of the hens’ low visual resolution on their inability to generalize responding across slide images and real conspecifics are discussed