To realize a stable hygric indoor climate, hygroscopic moisture buffering by materials plays an important role. Not only finishing materials on indoor surfaces, but also textiles (curtains, pillows) and paper (books) show an important moisture buffering capacity. In this paper, the role of books in the stabilization of indoor climates is studied. Recently, a model for hygroscopic buffering by books has been presented (Derluyn et al. 2007) introducing an effective vapour permeability and moisture capacity. This model is applied for the modelling of small scale experiments performed on book samples of 40x40x20 mm. The samples are initially in equilibrium at 0 % RH and then subsequently exposed to a step in RH to 54% and 79.5% RH. In a next step the relative humidity decreases again to 54% RH and finally, a cyclic RH is imposed, fluctuating between 54% and 79.5% RH. During the test, the weight change of the samples is monitored. As paper clearly behaves hysteretic, a hygric hysteresis model is incorporated based on the independent domains theory. A good agreement between measurement and simulation is obtained. In a second experiment, the hygroscopic buffering of books on a shelf in a room is measured. First the climatic room is conditioned at constant temperature and relative humidity. The books are in equilibrium with the relative humidity of the room. Then, a moisture production protocol is imposed and the relative humidity in the room is monitored. A fan helps to obtain well mixed air conditions. Air speed, air temperature and relative humidity is measured at different positions in the room and around the books, specifically near to the book shelf and in between the book shelves. The weight change of the books is measured. The results are compared with a room without books on the shelves. The moisture balance for the room is modeled using two different approaches. In the first approach, well mixed air conditions are assumed, while in the second approach the air flow around the books is modeled using CFD and location dependent mass surface coefficients are determined. The simulation results for the two approaches are compared with the measurements.