Degrees of sleep apnoea and daytime sleepiness are quite common in community populations. However the relationship between the two is poor, although sleepiness does correlate better with a history of snoring. It has been suggested that sleep can be fragmented by upper airways obstructive events, short of full apnoeas or hypopnoeas, and that these events may not provoke full cortical arousal, but be detectable through activation of the autonomic system. Failure to detect both these could mask a relationship between 'sleep apnoea' and daytime sleepiness. We have therefore measured sleepiness (Epworth Sleepiness Scale) in addition to both autonomic 'arousals' and inspiratory effort (using pulse transit time) in 473 men and women at home. Although sleepiness was related to a history of snoring, it was not significantly predicted by the measures of autonomic 'arousal', or inspiratory effort. Reported snoring and objectively measured snoring correlated poorly. As in other studies, nocturnal hypoxic dips were correlated with obesity, age, alcohol consumption, drug usage and a history of snoring. These data make it unlikely that sleep fragmentation from subtle variants of sleep apnoea and 'autonomic' (or 'subcortical') arousals are an important source of daytime sleepiness in the community.