The disruption of circadian rhythms following time-zone transitions gives rise to the syndrome of jet lag. The power of some of the symptoms of jet lag to predict the amount of jet lag measured at the same and at different times of the day has been investigated. Eleven healthy subjects were studied in an Isolation Unit for two days after a simulated flight from the UK to Beijing (8 time zones to the east). At six time-points (08:30, 11:00, 14:00, 17:00, 20:00, and 23:00 h), the subjects recorded their jet lag, and the differences from "normal" (that is, from days in which there is no jet lag) of alertness, hunger, indigestion, concentration, motivation, and irritability. They recorded at 08:30 h the type of food they had eaten since rising at 08:00h and, at the other times, the type of food eaten in the last three hours. Assessments were made by visual analogue scales or, in the case of type of food, by a nominal scale. Following the time-zone transition, the adjustment of meals appeared to be complete almost immediately. Jet lag and its symptoms were present during both experimental days. Jet lag tended to rise during the course of the daytime, accompanied by falls in alertness, motivation, and concentration. Correlation matrices between jet lag and each of the other variables were produced, using lags between the variable (from up to 5 time-points before the assessment of jet lag to 5 time-points afterwards) and pooling the results from both days. These matrices indicated that significant correlations existed only between jet lag and alertness, concentration, and motivation, and then only when these other variables were assessed at the same time as jet lag or 1 or 2 time-points earlier. Jet lag was then treated as the dependent variable and the symptoms as covariates in analysis of covariances (ANCOVAs), with the days treated as a random effect. This analysis enabled the significance of potential predictors of jet lag, together with their beta-coefficients (the relationship between a unit change of each significant predictor and the change in jet lag), to be calculated. Falls in alertness and motivation were significant predictors of increased jet lag, provided that they were measured at the same time, when they accounted for about 50% of the jet lag; when measured at other time-points, they did not act as significant predictors. It is concluded that the amount of jet lag varies during the course of the day and that it can be predicted from contemporaneous assessments of alertness and motivation-but not from assessments made at other times of the day, nor from other variables that are symptoms of jet lag, even though these other variables are significantly increased. In considering the results of this and our previous study, we reiterate the view that the exact meaning of "jet lag" is complex and that the particular combination of factors that contribute to it might vary with the time of day that the assessment is made. Inferences about any decrements due to time-zone transitions cannot be made reliably at times of the day that differ from the time when jet lag is assessed.