Researchers at Birmingham University have found what they described as a strong link between childhood cancers and exhaust emissions.
Postal addresses of 22,500 children who had died from cancer in Britain between 1955 and 1980 were compared with existing maps of atmospheric pollution levels.
Professor George Knox found these deaths were linked to hotspots for specific chemicals which included carbon monoxide, nitrogen oxides, benzene, dioxins and volatile organic compounds.
He has also called for such emissions to be brought under tighter controls.
Emission sources such as hospitals, bus/train stations, heavy transport hubs, and oil installations, were located using maps and information from the National Atmospheric Emissions Inventory.
Carbon monoxide and 1,3-butadiene, which are both produced by vehicle exhausts - in particular diesel engines - were among the primary culprits.
Prof Knox, emeritus professor of epidemiology and public health, said: "Although most cars are now fitted with catalytic converters as standard, most lorries, trucks and buses are not.
"And while there has also been a major increase in the levels of domestic traffic,
many of the country's motorways had not been built during the examined period.
"There are a lot of emission hotspots in urban areas like Birmingham, London, Glasgow and Manchester. These places have their clusters where cases occur."
He plotted the number of predicted deaths against actual deaths from childhood cancers, then the postcodes - where they were born, had lived and died - were used to calculate distances from the particular zones and emissions sources.
These calculations revealed an excess risk of cancer for children living within 300 metres of a chemical emissions hot spot and within one kilometre of an emissions source, such as a transport hub.
However a children's cancer charity yesterday criticised the professor's findings as "out of date".
Geoff Thaxter, director of services for CLIC Sargent, said: "We want to see conclusive research, but not reports which simply raise alarm without offering solutions to the public.
"We're not saying it's rubbish but it's looking at out of date statistics and there's no mention of a causal link."
Mr Thaxter also criticised the British Journal of Epidemiology and Community Health, which published Professor Knox's study today.
Prof Knox's report reveals that when living close to a source such as a bus or coach station, a child was at 12 times more likely to die from cancer. He suggests exposing a child in the womb, and soon after birth, to atmospheric pollutants is likely to be the critical period.
Prof Knox added that accepted safety levels for 1,3-butadiene in the workplace are unlikely to protect babies from developing cancer.