Why does so much hospital infection still exist despite the attention that it attracts? There were about 9·6 million UK National Health Service (NHS) hospital admissions in 2002–03. 5–10% of inpatients acquire an infection during their stay, and thus about 0·5–1·0 million hospital infections occur each year in the UK alone. Graham Ayliffe and Mary English offer a fascinating account of how the relation between hospitals and infection has altered through the centuries, mainly because of the realisation that many diseases were not a result of bad air (miasma). Although we now understand routes of cross-infection and have preventive strategies for most of these, understanding how beliefs and approaches to hospital infection have changed is instructive for challenges that still exist. In Roman times, hospital orderlies were known as nosocomi, the word deriving from Latin and Greek words for hospital. It is appropriate that in modern times nosocomial (meaning hospital acquired) is so closely linked to an old name for health-care workers. Health-care workers remain inadvertently responsible for most cases of cross-infection, either directly or indirectly via their hands. Airborne transmission of infection is now an uncommon cause of most hospital-acquired infections, but miasmas were widely espoused from the 6th to the 19th centuries as the way in which disease was spread. Ironically, however, there are examples of infections that may be transmitted by aerosols, which continue to confound us. Why have hospitals been over-run by outbreaks of viral gastroenteritis for much of the past 2years? How did severe acute respiratory syndrome spread so dramatically in modern well equipped hospitals? How would we cope with new cases of smallpox? As Ayliffe and English point out, it is still unknown how some of the last cases of hospital-acquired smallpox were contracted. The prevention and control of hospital infection must be underpinned by surveillance, which was pioneered in the USA, notably by the work of Robert Haley in the 1970s. These studies showed that intensive surveillance, feedback to clinicians, and a control prog-ramme could cut rates of hospital infection by about a third. Clinician feedback is fundamental to sharing the responsibility for controlling hospital infection. The US Communicable Disease Center established the National Nosocomial Infections Scheme, which runs to this day and has been copied in many countries. The UK Department of Health has recognised that standardised surveillance methods that are affordable and yet based on Haley's principles are one way forward. These allow comparison of infection rates or proxy measurements between units, but clearly case mix and unit type must be borne in mind; meticillin-resistant Staphylococcus aureus (MRSA) bacteraemia rates are now one of the performance indicators for NHS trusts. This book has many evocative photographs and illustrations. One of these, a Nightingale ward in 1900, reminds me of the picture on my office wall. My grandmother and other nursing staff replete in their starched uniforms are grouped around the ward. Both photographs ooze cleanliness and order, but both are filled with empty beds. The high number and turnover of patients is a major modern day obstacle to ensuring good infection-control practice. Ayliffe and English note that persuading staff to do the basics now seems to be as difficult as it was for Ignaz Semmelweis. Unfortunately, Semmelweis, dismayed by the reluctance of his peers to believe his exhortations about the need to wash their hands to control the spread of puerperal sepsis, was eventually committed to a sanatorium in Vienna, where he died of septicaemia in 1865. Shortly after Semmelweis died, Joseph Lister's first paper on antisepsis was published. In 1999, a UK NHS Executive Health Service Circular instructed health-care providers to strengthen the prevention and control of communicable disease and infection-control processes. Specifically they were asked to “put infection control and basic hygiene where they belong, at the heart of good management and clinical practice with appropriate resources”: fine words, but actions speak louder than words.