Against the background of various other theories, the well-established D'Arcy/Watt model is applied to sorption data of wool between 20 and 100 degreesC to determine its suitability to describe the isotherms and to systematize their temperature dependence. The model contains three components that represent two types of primary adsorption processes (Langmuir- and Henry-type adsorption) and a third one, describing multilayer formation of water molecules. Sorption isotherm data, as taken from the literature, could, in all cases, be fitted extremely well by the model. The temperature dependence of the five parameters of the model, related to the continuous decrease of water regain with increasing temperature for all humidities, reveals a number of inconsistencies. Probably the most important of these is that the Langmuir capacity constant apparently becomes zero at the glass transition temperature of dry wool. This is at variance with the idea of specific molecular sites of water sorption, which is inherent to the model. Other inconsistencies relate to the small van't Hoff enthalpies and possible compensation effects for various parameters. These observations indicate that the D'Arcy/Watt theory, despite its physicochemical plausibility and empirical success, overinterprets the complexity of the mechanisms underlying the sorption behavior of wool and other cu-keratin fibers.