Much of the extensive work on hospital cross-infection has of necessity been opportunist associated with the occurrence of sporadic outbreaks of infection. Investigations concentrated on such outbreaks cannot give a true picture of the general problem, and the value of any counter-measures which may have been introduced is not easily assessed.A pilot study, carried out in a large dermatology department, revealed an ideal model for the planned, long-term investigation of hospital infection (Selwyn, 1963). An endemic situation was uncovered in which the extensive skin lesions of in-patients were readily colonized by the abundant pathogenic bacteria present in the environment. Many of the infected patients became in their turn prolific dispersers of pathogens. A detailed study of these processes was facilitated by the long periods of hospital treatment which many patients with skin diseases require. Further work was therefore carried out in this neglected field to investigate the dynamic relationship which exists between the bacterial flora of these patients and that of their environment. The effect of simple anti-bacterial measures upon this relationship was also studied during two of the four stages of the investigation. The principal aim here was to assess the relative importance of nasal carriers and patients with infected lesions in the aetiology of hospital infection.
MATERIALS AND METHODS
Sampling proceduresThe investigation was carried out over a 2-year period in three dermatological wards. These are housed in a relatively modern pavilion, within a large teaching hospital. The two main wards, situated on different floors, are 68 ft. long, 28 ft. wide and 12 ft. high; each contains sixteen beds. The third ward, containing eight beds, is 32 ft. long; the width and height are as in the main wards.The degree of bacterial contamination was measured on two days of the week at representative sites in each ward, and in the adjacent day-room and bathroom. Air counts were performed with pairs of 3-5 in. Petri dishes containing blood agar and 7 5 % salt-milk agar respectively. These were exposed for 2 hr. periods as 'settle plates' and during each period similar plates were exposed in a 'Casella' slit sampler (but this could not be operated in the bathrooms). Solid surfaces were examined by contact-transfer using self-adhesive cellulose tape in a similar manner to that used by Thomas (1961) for swabbing skin. Direct impressions