Malaria transmission is known to be strongly impacted by temperature. The current understanding of how temperature affects mosquito and parasite life history traits derives from a limited number of empirical studies. These studies, some dating back to the early part of last century, are often poorly controlled, have limited replication, explore a narrow range of temperatures, and use a mixture of parasite and mosquito species. Here, we use a single pairing of the Asian mosquito vector, An. stephensi and the human malaria parasite, P. falciparum to conduct a comprehensive evaluation of the thermal performance curves of a range of mosquito and parasite traits relevant to transmission. We show that biting rate, adult mortality rate, parasite development rate, and vector competence are temperature sensitive. Importantly, we find qualitative and quantitative differences to the assumed temperature-dependent relationships. To explore the overall implications of temperature for transmission, we first use a standard model of relative vectorial capacity. This approach suggests a temperature optimum for transmission of 29°C, with minimum and maximum temperatures of 12°C and 38°C, respectively. However, the robustness of the vectorial capacity approach is challenged by the fact that the empirical data violate several of the model's simplifying assumptions. Accordingly, we present an alternative model of relative force of infection that better captures the observed biology of the vector-parasite interaction. This model suggests a temperature optimum for transmission of 26°C, with a minimum and maximum of 17°C and 35°C, respectively. The differences between the models lead to potentially divergent predictions for the potential impacts of current and future climate change on malaria transmission. The study provides a framework for more detailed, system-specific studies that are essential to develop an improved understanding on the effects of temperature on malaria transmission.