A cross-system, worldwide approach has been used to ascertain the spatial, seasonal and long-term variability of areal phytoplankton photosynthesis (PP) in lakes using published data sets. Also, the average fraction of annual PP occurring under ice is calculated. The lakes considered embrace a range of properties (depth. mixing, flushing rate, latitude and trophic status). The overall yearly PP distribution is skewed to the left, suggesting the dominance of low PP rates in the data set. When comparing lake types, no differences in average PP have been found among them. In particular, there are no clear areal PP differences among lakes of different trophic status on yearly, averaged basis, suggesting that environmental limitations to PP also exist in lakes of higher trophic status. Volumetric-based PP can be better used to outline PP-based trophic differences, but some degree of overlap is also apparent. Across ail lake types (except in tropical lakes), the PP seasonal course experiences only one peak in the year, but its timing is clearly different for each lake type. The seasonal variability of PP is lower in tropical lakes, as previously reported, but the variability of the other lake types is roughly the same. Therefore, the effects of depth, mixing regime, flushing rate and nutrient status on PP seasonality are difficult to ascertain since they appear to be counterbalanced by other more pervasive. local effects. Particularly, there is no increase in temporal variability with the trophic status of lakes, suggesting that PP seasonal control by physical variables overrides that of nutrients. Also, no significant relationship between average PP and latitude has been found. Seasonal variability increases as the yearly PP increases. On a relative basis, there is a spatial gradient of seasonal variability of PP, which is weaker when seasonal variability of PP is considered in interyear comparisons. Long-term (i.e. interannual) variability of PP is clearly related to increasing yearly averaged PP. Specifically, in temperate, stratifying lakes the seasonal time course of PP is clearly different from that of phytoplankton biomass, suggesting an uncoupling of both variables as a result of differing P-max and losses throughout the year. On an average basis, environmental variables are poor predictors of areal daily PP, thereby implying that the interplay of factors is complex and changing throughout the year. PP under ice averages 10% of yearly PP, but its variability is high enough to make its measurement advisable.