They say it may explain why almost all children who have peanut allergy have an eczema rash in their first six months of life. It's suggested the allergens get in through the skin of babies with eczema, sensitising them to peanuts and increasing the risk of an allergy.

"We acknowledge that kissing is essential for a child's social and emotional development and do not advocate avoidance of kissing, but caution against high-saliva-volume kisses soon after peanut consumption in infants with eczema, '' they say in a letter in the Journal of Allergy and Clinical Immunology.

Peanut allergy is usually diagnosed in children under two, and in the past 20 years the number of children diagnosed has nearly doubled. Just how and why some children develop the allergy and others do not is not clear. What is known is that in some babies, exposure to peanuts leads to sensitisation, which means the child develops antibodies to the allergen. Most children who become sensitised do not develop a problem allergy, but around one-in-seven do.

In their letter, the specialists from the Alfred Hospital in Melbourne, Monash University and the Princess Margaret Hospital for Children in Perth, say that saliva containing traces of peanuts may be implicated.

"We hypothesise that cutaneous exposure to peanutcontaining saliva in infants with eczema might induce allergic sensitisation to peanut, '' they say.

They say that up to 12ml of saliva can be transferred to a baby's forehead with a "big, affectionate kiss", and that a single kiss could transfer up to 88.8mg of peanut proteins in saliva.

"Interestingly, families of children who become allergic to peanut tend to consume more peanut than families with children without peanut allergy. Almost all children who have peanut allergy have an 'oozing, crusted rash' during their first six months of life, '' says the report.

Dr George Du Toit, paediatric allergy consultant at the Evelina Children's Hospital, Guy's and St Thomas' NHS Trust, who is involved in a major new UK study, said the Australian study was interesting but that kissing was only one of many possible routes of environmental peanut exposure.

"Individuals with allergies generally tend to have had eczema as babies and 20% of children in the UK develop eczematous rashes, '' he said. "Most peanut-induced allergic reactions occur during the first three years of life and usually on first known ingestion of peanut.

This implies that ingesting peanut is not always the route of sensitisation.

"Kissing is one possible route of sensitisation, but there are other important mechanisms such as handling of the infant with peanut-contaminated hands. For example, a mother who prepares a peanut butter sandwich is likely to handle her young infant during the preparation and consumption thereby transferring peanut to the skin of the baby. It has also been shown that traces of peanut may still be found on hands and table surfaces even after washing. In my opinion, the epidemic of peanut allergy would be difficult to blame on peanut exposure through kissing alone.''