Soft bedding has been shown in epidemiologic studies to increase the risk for sudden, unexpected death in prone-sleeping infants. We compared the physical properties of conventional bedding to bedding from two sources: 1) bedding that covered the airways of victims of sudden infant death syndrome (SIDS) lying prone and face down at the time of death; and, 2) bedding associated with increased risk for SIDS in case-control studies (i.e. bedding filled with ti tree bark). Using simple mechanical models and the head from an infant mannequin, we measured the resistance to airflow, malleability, and capacity to limit CO2 dispersal of the bedding. We also describe a technique for quantifying bedding softness. The resistance and malleability were similar for the conventional bedding, the ti tree bedding, and the bedding from SIDS deaths (analysis of variance, p = 0.85 and 0.16). The ti tree bedding and the other bedding from SIDS cases differed from conventional bedding in two physical properties. Both groups were softer (p < or = 0.005) and limited CO2 dispersal to a greater degree (p < or = 0.009). The finding that increased capacity to limit CO2 dispersal is a consistent property of the bedding covering the airways of these SIDS victims and of bedding shown to be an epidemiologic risk factor for SIDS supports rebreathing of expired air as a mechanism underlying the association of certain kinds of bedding with SIDS.