Two unpublished letters of Robert Louis Stevenson

This article prints and discusses two uncollected letters by Robert Louis Stevenson to the publishing firm of Chatto & Windus, concerning his book Memories and Portraits.

Diviser, choisir et condenser: rubriques et illustrations du règne d'Uter et Pandragon dans le Merlin de Robert de Boron

La sélection opérée par les rubriques et les enluminures, formes contraintes qui échappent dans une certaine mesure à la linéarité du développement textuel, détermine la réception de l'œuvre. Les manuscrits comprenant des miniatures synthétiques (comme celles qui représentent l'enfance de Merlin) ou des séries iconographiques (comme celles qui rendent compte des amours d'Uterpandragon et Ygerne et de leurs conséquences) présentent donc un caractère exceptionnel, orientant la lecture du texte et son interprétation.

Construction impossible et défense improbable : la tour du roi Vertigier dans l'Historia Brittonum de Nennius, l'Historia Regum Britanniae de Geoffroy de Monmouth, le Brut de Wace et le Merlin de Robert de Boron

La question architecturale de l'impossibilité de l'usurpateur Vertigier à construire sa tour se double d'une réflexion politique et morale déterminante. La tour branlante est en effet à l'image du royaume déchiré de Grande Bretagne : elle emblématise l'instabilité politique et les affrontements militaires qui déterminent l'ascension illégitime puis la chute de ce mauvais roi.

Empirical test of an agricultural landscape model: the importance of farmer preference for risk aversion and crop complexity

Developing models to predict the effects of social and economic change on agricultural landscapes is an important challenge. Model development often involves making decisions about which aspects of the system require detailed description and which are reasonably insensitive to the assumptions. However, important components of the system are often left out because parameter estimates are unavailable. In particular, measurements of the relative influence of different objectives, such as risk, environmental management, on farmer decision making, have proven difficult to quantify. We describe a model that can make predictions of land use on the basis of profit alone or with the inclusion of explicit additional objectives. Importantly, our model is specifically designed to use parameter estimates for additional objectives obtained via farmer interviews. By statistically comparing the outputs of this model with a large farm-level land-use data set, we show that cropping patterns in the United Kingdom contain a significant contribution from farmer’s preference for objectives other than profit. In particular, we found that risk aversion had an effect on the accuracy of model predictions, whereas preference for a particular number of crops grown was less important. While nonprofit objectives have frequently been identified as factors in farmers’ decision making, our results take this analysis further by demonstrating the relationship between these preferences and actual cropping patterns.