Abstract: Herbivores challenged by diets with high concentrations of tannins learn by individual experience to self-select medicinal compounds such as polyethylene glycol (PEG), which neutralizes the negative postingestive effects of tannins. We investigated the transmission of this acquired self-medicative behaviour from mother to offspring. One group of ewes (experienced, N = 8) was conditioned to associate the beneficial effects of PEG after consuming a tannin-rich diet. Ewes ingested a meal of high-tannin food and were then offered PEG. Subsequently, ewes ingested the same tannin-rich meal and were then offered a food (grape pomace; control) that did not have the medicinal effects of PEG. After conditioning, the experienced group and a naïve group of ewes (N = 8) were given a choice between the high-tannin food, PEG and grape pomace. Experienced ewes showed higher intake and preference for PEG than did naïve ewes (P < 0.05). Subsequently, experienced and naïve ewes with their naïve lambs, as well as a group of naïve lambs without their mothers (N = 8), were exposed to the tannin-rich diet, PEG and grape pomace. Lambs were then tested for their ability to self-medicate with PEG by offering them a choice between the tannin-rich diet, PEG and grape pomace. Lambs from experienced and naïve mothers showed a higher preference for PEG than did lambs exposed without their mothers (P = 0.05). Thus, the presence of the mother (experienced or naïve) was important for naïve lambs to learn about the medicinal benefits of PEG. We conclude that the mother's presence per se may increase the efficiency of creating new knowledge, such as preference for a medicine, within a group, beyond transmitting and maintaining this knowledge across generations.

Abstract: In search of the minimal requirements for transitive reasoning, a simple neural network was trained and tested on the non-verbal version of the conventional “five-term-series task” – a paradigm used with human adults, children and a variety of non-human species. The transitive performance of the network was analogous in several aspects to that reported for children and animals. The three effects usually associated with transitive choices i.e. “symbolic distance”, “lexical marking” and “end-anchor”, were also clearly shown by the neural network. In a second experiment, where the training conditions were manipulated, the network failed to match the behavioural pattern reported for human adults in the test following an ordered presentation of the premises. However, it mimicked young children's performance when tested with a novel comparison term. Although we do not intend to suggest a new model of transitive inference, we conclude, in line with other authors, that a simple error-correcting rule can generate transitive behaviour similar to the choice pattern of children and animals in the binary form of the five-term-series task without requiring high-order logical or paralogical abilities. The analysis of the training history and of the final internal structure of the network reveals the associative strategy employed. However, our results indicate that the scope of the associative strategy used by the network might be limited. The extent to which the conventional five-term-series task, in absence of appropriate manipulations of training and testing conditions, is suitable to detect cognitive differences across species is also discussed on the basis of our results.

Abstract: Value transfer theory has been proposed to account for transitive inference effects (L. V. Fersen, C. D. L. Wynne, J. D. Delius, & J. E. R. Staddon, 1991), in which following training on 4 simultaneous discriminations (A+B-, B+C-, C+D-, D+E-) pigeons show a preference for B over D. According to this theory, some of the value of reinforcement acquired by each S+ transfers to the S-. In the transitive inference experiment, C (associated with both reward and nonreward) can transfer less value to D than A (associated only with reward) can transfer to B. Support for value transfer theory was demonstrated in 2 experiments in which an S- presented in the context of a stimulus to which responses were always reinforced (S+) was preferred over an S- presented in the context of a stimulus to which responses were sometimes reinforced (S +/-).

Abstract: Pigeons were trained successively either on 2 delayed simple discriminations or on a delayed simple discrimination followed by delayed matching-to-sample. During subsequent transfer tests, the initial stimuli from the 1st task were substituted for those in the 2nd. Performances transferred immediately if both sets of initial stimuli had been associated with the presence versus absence of food on their respective retention tests, and the direction of transfer (positive or negative) depended on whether the substitution involved stimuli with identical or different outcome associates. No transfer was found, however, when the initial stimuli were associated with different patterns of responding but food occurred at the end of every trial. These results are consistent with outcome expectancy mediation but are incompatible with response intention and retrospective coding accounts.

Abstract: Soiled bedding influences cleanliness and disease levels in dairy cows and there is no evidence of an inherent latrine behaviour in cattle. If cows were trained to use a concrete area of the housing system as a latrine, a cleaner bed could be maintained. Thirteen group-housed, 14-16-month-old Holstein-Friesian heifers, were clicker trained with heifer-rearing concentrate pellets as a reward. Training was carried out in four phases. (Phase 1) Association of feed reward with clicker, criterion: 34/40 correct responses. (Phase 2) Simple task (nose-butting a disc) to reinforce phase 1 association, criterion: 17/20 correct responses. (Phase 3) Association of eliminative behaviour with reward where criterion was four sessions with only one incorrect response: criteria for each heifer in phases 1-3 were set using binomial tests. (Phase 4) Shaping eliminative behaviour to occur on concrete. Possible responses were, eliminating on concrete (C) or straw (S), or moving from one substrate to another immediately before eliminating: C --> S, S --> C. Heifers were rewarded for the desired behaviours C and S --> C and ignored when S and C --> S occurred. If learning was achieved, C should increase as C --> S decreased and S --> C should increase as S decreased: tested with Spearman rank correlations. All heifers achieved criterion by day 4 of phase 1 (P = 0.001); day 1 of phase 2 (P = 0.001) and day 10 of phase 3 (P < 0.009). Responses changed throughout phase 3 beginning with (i) looking at the trainer whilst voiding then moving to trainer after the click, and later including (ii) moving to trainer immediately before- or (iii) during voiding. No relationship was found between S and S --> C (rs = -0.14; P = 0.63) or C and C --> S (rs = -0.33; P = 0.25). All group members eliminated more often on concrete (580) than on straw (141) but four heifers with consistently longer lying bouts also showed more C --> S before lying down (Mann-Whitney, P = 0.007). The present study is believed to be the first reported work to show that cattle can be trained to show an awareness of their own eliminative behaviour. This was not successfully shaped to latrine behaviour, however, and it is suggested that floor type may not have been a sufficiently salient cue. Voiding on straw occurred largely with response C --> S (0.73) and general behaviour suggested that this was strongly linked to lying patterns of individual heifers.