This paper documents a significant effect of short-term temperature fluctuations on attitudes towards institutions and on civil unrest in Africa. Combining attitudinal survey and climate data, we calculate temperature as perceived by respondents via an algorithm that combines different meteorological variables. The results show that daily temperature anomalies at the location of interview increase self-reported mistrust in government and intentions to vote for opposition parties. Effects are particularly strong in poor countries where temperature anomalies also increase self- reported intentions to protest. Accordingly, we find that temperature anomalies also increase incidences of protests and riots. Evidence suggests that effects are not driven by changes in agricultural incomes.