Preventing cot death is the aim of a Pounds 450,000 specialist unit established at Bristol University's Institute of Child Health.
Supported by The Foundation for Study of Infant Death and the Anne Diamond Cot Death Appeal, the unit aims to build on research already done at Bristol by Peter Flemming, Jean Golding and Jem Berry.
"Back in the late 1980s we identified the major risk factors," explained Professor Berry. "They are parental smoking, overwrapping and allowing the baby to sleep on its stomach."
These findings resulted in an education campaign among parents in Avon, which led to a massive reduction in cot deaths, from 40 a year in the late 1980s to an average of six per year now.
Avon's encouraging results led to a nationwide campaign which succeeded in halving cot deaths in Britain. Back in the 1980s, an average of 2,000 babies died of the syndrome each year. That figure is now down to less than 1,000.
But more research is necessary because cot death remains the commonest cause of death among babies aged between one month and one year. So, in a bid to reduce it further, the new unit is evaluating how families care for their babies.
A Department of Health-funded investigation is attempting to discover whether babies who die of cot death syndrome are treated any differently from other babies.
Parents who have lost a child in this way will be questioned about the baby's sleeping environment and sleeping, feeding, and medical history as well as their own smoking habits.
Each case will then be compared with four other families whose babies are healthy, to establish whether there are any significant differences in methods of care.
This programme will be backed up by physiological tests in Bristol's sleep laboratory, where babies' breathing, temperature, heart rate and oxygen intake will be recorded as they sleep.
The Bristol unit is one of two established this summer by the Foundation for the Study of Infant Deaths. The second, a Pounds 250,000 centre at Liverpool will focus both on care after birth and on foetal development beforehand.