There have been several attempts in recent years to relate the incidence of sickness in the Royal Navy to variations of the thermal environment of ships' companies. During the second world war the incidence of the common causes of ill health, as revealed by attendances at the sick bay, was shown to be nearly twice as great in the ships of the Eastern and British Pacific Fleets as it was in shore establish ments, where the mean air temperature was usually about 7?F. lower than the temperature between decks; this difference was due largely to an increase in the incidence of skin disease (Ellis, 1947, 1948). Time lost due to sickness was doubled or trebled when ships joined the Eastern Fleet from the Mediterranean or Home Fleets. When the war ended, relaxation of security precautions concerning the movement of warships made it possible for such information to be included in the monthly returns of sickness to the Admiralty. These returns revealed that during the first year after the war the relative sickness incidence was more than doubled in tropical waters as contrasted with northern temperate waters, and diseases of the skin were increased by three to four times (Roberts, 1948). Recent work suggests that for different classes of ships it may be possible to predict the mean mess deck temperature with reasonable accuracy provided the mean upper deck temperature is known (Gray and Smith, 1952). In this event it may be possible to relate sickness to the average thermal environment between decks more accurately than has been pos sible hitherto. In 1948 it was requested that the mean upper deck temperature at noon should be included with each monthly return. The results of the analysis of the returns including this information for the year October 1, 1948, to September 30, 1949, are described below. Methods