Domestic dogs (Canis lupus familiaris) were tested in four experiments for their understanding of means-end connections. In each of the experiments, the dogs attempted to retrieve a food treat that could be seen behind a barrier and which was connected, via string, to a within-reach wooden block. In the experiments, either one or two strings were present, but the treat was attached only to one string. Successful retrieval of the treat required the animals to pull the appropriate string (either by pawing or by grasping the wooden block in their jaws) until the treat emerged from under the barrier. The results showed that the dogs were successful if the treat was in a perpendicular line to the barrier, i.e. straight ahead, but not when the string was at an angle: in the latter condition, the typical response was a proximity error in that the dogs pawed or mouthed at a location closest in line to the treat. When two strings that crossed were present, the dogs tended to pull on the wrong string. The combined results from the experiments show that, although dogs can learn to pull on a string to obtain food, they do not spontaneously understand means-end connections involving strings.

We conducted four experiments with 56 adult dogs (Canis lupus familiaris) involving tasks where food was dropped through an opaque tube connected either vertically or diagonally to one of two or three goal boxes. In the first experiment, modelled after studies with children and primates, the dogs first searched significantly more often in the location directly beneath the drop-off point (a gravity bias), although this box was not connected with the tube. These results are comparable to those of human infants and cotton-top tamarins. Experiments 2-4 tested which problem solving strategy the dogs applied to find the food. Results show that they do not understand the physical mechanism of the tube itself, and they apply one of three search strategies: search the gravity box (the one below the drop-off box); search the box in the middle; learn the correct location of the goal box. When the goal box was in the same location the dogs learned to search there over trials, that is, they learned to 'defy gravity', but when the location of the goal box changed over trials they showed no learning. These findings are compared with those from human infants and cotton-top tamarins: like these species, the dogs can learn to overcome a gravity bias, but only when the reward is to be found in a consistent location.

Pigeons (Columba livia), gray squirrels (Sciurus carolinensis), and undergraduates (Homo sapiens) learned discrimination tasks involving multiple mutually redundant dimensions. First, pigeons and undergraduates learned conditional discriminations between stimuli composed of three spatially separated dimensions, after first learning to discriminate the individual elements of the stimuli. When subsequently tested with stimuli in which one of the dimensions took an anomalous value, the majority of both species categorized test stimuli by their overall similarity to training stimuli. However some individuals of both species categorized them according to a single dimension. In a second set of experiments, squirrels, pigeons, and undergraduates learned go/no-go discriminations using multiple simultaneous presentations of stimuli composed of three spatially integrated, highly salient dimensions. The tendency to categorize test stimuli including anomalous dimension values unidimensionally was higher than in the first set of experiments and did not differ significantly between species. The authors conclude that unidimensional categorization of multidimensional stimuli is not diagnostic for analytic cognitive processing, and that any differences between human's and pigeons' behavior in such tasks are not due to special features of avian visual cognition.

We investigated perseveration and detour behaviour in 36 equids (Equus caballus, E. asinus, E. caballus × E. asinus) and compared these data to those of a previous study on domestic dogs (Canis familiaris). The animals were required to make a detour through a gap at one end of a straight barrier in order to reach a visible target. After one, two, three or four repeats (A trials), the gap was moved to the opposite end of the barrier (B trials). We recorded initial deviations from the correct solution path and the latency to crossing the barrier. In the A trials, mules crossed the barrier significantly faster than their parental species, the horses and donkeys. In the B trials, following the change of gap location, all species showed a reduction in performance. Both dogs and horses exhibited significant spatial perseveration, going initially to the previous gap location. Donkeys and mules, however, performed at chance level. Our results suggest that hybrid vigour in mules extends to spatial abilities.

This study compares the behaviour of the mule (Equus asinus x Equus caballus) with that of its parent species to assess the effects of hybridization on cognition. Six mules, six ponies (E. caballus) and six donkeys (E. asinus) were given a two choice visual discrimination learning task. Each session consisted of 12 trials and pass level was reached when subjects chose the correct stimulus for at least 9 out of the 12 trials in three consecutive sessions. A record was made of how many pairs each subject learnt over 25 sessions. The mules' performance was significantly better than that of either of the parent species (Kruskal-Wallis: H(x) = 8.11, P = 0.017). They were also the only group to learn enough pairs to be able to show a successive reduction in the number of sessions required to reach criterion level. This study provides the first empirical evidence that the improved characteristics of mules may be extended from physical attributes to cognitive function.