The reliability of spirometry is dependent on strict quality control. We examined whether quality control criteria recommended for adults could be applied to children aged 2-5 years. Forty-two children with cystic fibrosis and 37 healthy children attempted spirometry during their first visit to our laboratory. Whereas 59 children (75%) were able to produce a technically satisfactory forced expiration lasting 0.5 second, only 46 (58%) could produce an expiration lasting 1 second, with the youngest children having the most difficulty. Start of test criteria for adults were inappropriate for this age group, with only 16 of 59 children producing a volume of back extrapolation as a proportion of forced vital capacity of less than 5%, whereas all but 4 could produce a volume of back extrapolation of 80 ml or less. More than 90% of children were able to produce a second forced vital capacity and a second forced expired volume in 0.75 second within 10% of their highest. Errors in the spirometry software resulted in inaccurate reporting of expiratory duration and inappropriate timed expired volumes in some children. We describe recommendations for modified start of test and repeatability criteria for this age group, and for improvements in software to facilitate better quality control.