Background. Many studies have reported that heatwaves increase mortality. However, it is not certain whether less pronounced rises in temperature also increase it. Such information might be important for predicting the impact of potential weather changes on mortality. We have assessed the relationship between daily mortality and moderate increases in weather temperature in Barcelona, Spain, following a time series approach.Methods. The study included the period from 1 January 1985 to 30 December 1989. For all the population resident in Barcelona, Spain, we considered the following daity data: total mortality, mortality of those >65 years, and cardiovascular and respiratory mortality. The meteorological variables were: minimum temperature, maximum temperature, dew point temperature and relative humidity. Several transfer function (ARIMA) models were estimated for the entire period and for both winters and summers separately.Results. We found that unusual periods of at least three consecutive days of increased weather temperature Increased mortality, independently of the V-shaped relationship also found. The occurrence of an unusual period increased total daily mortality by 2% on average (1.7% on summers) and by 2.6% In those over 65 (2% on summers). Cardiovascular mortality rose by 4.6% (4.2% on summers) and respiratory mortality by 21.6% (13.2% on summers). However, only those unusual periods with an excess temperature and humidity were associated with mortality increases.Conclusions. The unusual periods observed in the present study cannot be classified as heatwaves because the weather temperature never reached high values and most of them occurred during the winter. The association of unusual periods with mortality was stronger during winters than in summers, maybe because unusual winter periods showed a temperature deviation from the average twice that in summer or because humidity during unusual winter periods was higher than in summer. © 1995 International Epidemiological Association.