It has frequently been claimed that cycling in heavy traffic is unhealthy, more so than driving a car. To test thishypothesis, teams of two cyclists and two car drivers in two cars were equipped with personal air samplers whiledriving for 4 h on 2 different days in the morning traffic of Copenhagen. The air sample charcoal tubes were analysedŽ.Ž.for their benzene, toluene, ethylbenzene and xylene BTEX content and the air filters for particles total dust . The concentrations of particles and BTEX in the cabin of the cars were 2-4 times greater than in the cyclists’ breathingzone, the greatest difference being for BTEX. Therefore, even after taking the increased respiration rate of cyclistsinto consideration, car drivers seem to be more exposed to airborne pollution than cyclists. 2001 Elsevier ScienceB.V. All rights reserved

A series of field studies were carried out in London, UK, during 1999–2000 in which over 400 fine particle (PM2.5)personal exposure level measurements were taken for journeys in bicycle, bus, car and underground rail transportmicroenvironments. This was the first comprehensive PM2.5personal exposure study of transport users. Both a fixed-route multi-transport mode study and a study of cyclists’ commuter journeys were undertaken. Subsequent to thesefield studies regression modelling of possible influencing factors of these exposure levels was carried out. Meteorologicalvariables, traffic density, mode and route were considered; the relationships of personal exposure levels with fixed sitemonitor (FSM) concentrations, and of the FSM concentrations with the potential predictor variables, were alsoinvestigated. This analysis of the determinants of transport user exposure to PM2.5in London, UK, showed that windspeed had a significant influence on personal exposure levels, though explained only up to 20% of the variability of roadtransport user exposure levels. The occurrence of higher wind speeds was strongly associated with a decrease in personalexposure levels; a 1.5–2.0 fold difference in exposure level concentrations was estimated between the 10th and 90thpercentiles of wind speed. Route was a significant factor, whilst mode was not a significant factor in the street microenvironment (between bicycle, bus and car modes); models incorporating route and mode, as well as wind speed,explained approximately 35% of the variability in PM2.5exposure levels. Personal exposure levels were reasonably correlated with urban background FSM concentrations, for fixed-route road mode (bicycle, bus and car) exposure level concentrations, r ¼ 0:27 ( p50:01) and for commuter cyclists’ exposure level concentrations r ¼ 0:58 ( p50:01). # 2001Elsevier Science Ltd. All rights reserved