The ergonomic evaluation of mattresses is largely ignored in the current literature. This is somewhat surprising given their importance and the length of time spent using them. This study considers some ergonomic aspects of their design, including body contact pressure and subjective ratiugs of comfort. Subjects (12 females) found all of the mattresses tested to be significantly more comfortable than an incompressible wooden reference surface. However, no significant differences were found between mattress types, which included orthopaedic and normal designs.. Analysis of body contact pressures (measured at the shoulder, elbow, hip, knee and ankle) found few significant differences between experimental conditions. It is argued that limitations in the methodology may not take account of the change in surface area and anatomical sites of contact under different conditions of mattress compressibility. No significant asselciations were found between comfort ratings and peak body contact pressures. It seems likely that subjective ratings of mattress comfort are dependent on a wider set of factors than contact pressure alone, a finding reported elsewhere in studies of seating.