We show that sleep deprivation exerts strong negative effects on mothers' labour market performance. To isolate exogenous variations in maternal sleep, we exploit unique variations in child sleep disruption using a UK panel dataset that follows mother-child pairs through time. We find that sleeping one hour less per night on average significantly decreases maternal labour force participation, the number of hours worked, and household income. We identify one mechanism driving the effects, namely the influence of maternal sleep on selection into full-time versus part-time work. Increased schedule flexibility for mothers with sufficient tenure mitigates the negative effects of sleep deprivation.