The aim of this paper was to develop a protocol to study the interpretation of ambiguous stimuli in cats as a measure of welfare. Ten cats were trained to discriminate between a rewarded position (R) and an unrewarded one (U), as measured by the approach latency for each position. After discrimination, they were exposed to three ambiguous unrewarded positions (R-near, R/U-equidistant, U-near) distributed at intermediate points between R and U. Approach latency increased as increasing the distance from the rewarded position: latencies to approach R and R-near were significantly shorter than for R/U-equidistant, U-near and U. This protocol should be further studied to assess its effectiveness in highlighting differences according to the welfare level of individual cats.

We investigated the incidental encoding in working memory of event serial order in tasks in which the relevant dimension was visual appearance, spatial location, or visual-spatial conjunction. The participants (n = 60) were asked whether two sequences were identical based on the relevant dimension, and to ignore changes in the irrelevant dimension, that is the order of events. Changes in serial order impaired the performance when sequences were identical; this effect was more pronounced in spatial sequences. It is suggested that the order was incidentally encoded with the relevant information to the task in an earlier stage in the information processing, which explains a different pattern of serial order effect according to the relevant dimension. Although encoded, the serial order may not have affected the visual storage in working memory because it might have been kept in a distinct subcomponent rather than the one that stores the visual characteristic. Moreover, the order may have affected spatial storage because the maintenance of this dimension might be related to a rehearsal mechanism based on serial order of sequence. This conclusion qualifies models that admit the architecture of working memory based on the specificity of encoding and functional interaction between subcomponents of storage.

L'agglomération spatiale des firmes est un phénomène largement dé-battu en économie géographique. Cet article propose une nouvelle explication à ce phénomène : les entreprises mettent en avant la localisation comme un moyen de signaler un bien nouveau. Accepter la concurrence en se localisant au plus près des entreprises existantes permet de montrer aux consommateurs la valeur du produit vendu. La polarisation géographique, et par là même une plus grande concurrence, sert de faire-valoir à une innovation. On montre que la firme innovante se localise dans la même région que la firme concurrente uniquement si le produit qu'elle propose est suffisamment innovant. Le critère de raffinement "K-undefeated" assure l'unicité de l'équilibre semi-séparateur. La polarisation de l'activité est uniquement imputable à l'asymétrie d'information. Cette forte polarisation doit inciter les autorités supra-régionales à privilégier plutôt les infrastructures de désenclavement qui se traduisent par une forte baisse des coûts de transport. En effet, une faible baisse du coût de transport peut se traduire par une baisse du bien-être total.