Researchers at first thought this might be because social conditions had improved over the past 20 years.

But further research indicates that better health care in the first month after birth is responsible for the fall.

Babies born to lone parents are still a third more likely to die in their first year than those born to couples. Researchers say this is down to socio-economic factors.

The report's authors say the finding highlights a need for sustained support for lone mothers.

'Unacceptable difference'

Dr Margaret Whitehead of the King's Fund, and Frances Drever of the Office of National Statistics conducted the study, which is published in the British Medical Journal.

Their starting point was that "infant mortality has long been accepted as an important indicator of a population's health".

They said that evidence of a difference by social group in death rates is "regarded as particularly unacceptable and a spur to action".

They then examined statistics for all live births - 14.3 million - in the UK between 1975 and 1996.

These details were compared with figures for infants who died aged between 0 and 12 months - 135,800 - over the same period.

Likelihood of death

The researchers found that in 1975, a child was 79% more likely to die in its first year if it were born to a lone mother than if it were born to a couple.

By 1996 this difference in risk had fallen to 33%.

Socio-economic factors are thought to influence survival chances

However, when the researchers broke down their statistics, they found that children born to lone parents were as likely to die as children born to couples in the two lowest social groups.

They used the registrar general's system, which divides people into five social groups. I is the highest and V is the lowest.

"For each social class, infant mortality has more than halved since 1975," they say. "But the gap between the top and the bottom of the social scale has shown little narrowing."

Baby deaths occur later

Their findings also showed that in 1996 fewer of the infant deaths occurred during the period just after birth than in 1975.

This, combined with their other findings, suggested the fall in the death rate occurred because neonatal health care improved over that time.

The disparity between social groups was down to affluence affecting a parent's ability to care for their child, they said.

"Post-neonatal mortality is thought to be influenced to a much greater extent by parents' socio-economic circumstances, reflected in the cause of death," they said.