Some of the most sophisticated of all animal-built structures are the mounds of African termites of the subfamily Macrotermitinae, the fungus-growing termites. They have long been studied as fascinating textbook examples of thermoregulation or ventilation of animal buildings. However, little research has been designed to provide critical tests of these paradigms, derived from a very small number of original papers. Here I review results from recent studies on Macrotermes bellicosus that considered the interdependence of ambient temperature, thermoregulation, ventilation and mound architecture, and that question some of the fundamental paradigms of termite mounds. M. bellicosus achieves thermal homeostasis within the mound, but ambient temperature has an influence too. In colonies in comparably cool habitats, mound architecture is adapted to reduce the loss of metabolically produced heat to the environment. While this has no negative consequences in small colonies, it produces a trade-off with gas exchange in large colonies, resulting in suboptimally low nest temperatures and increased CO(2) concentrations. Along with the alteration in mound architecture, the gas exchange/ventilation mechanism also changes. While mounds in the thermally appropriate savannah have a very efficient circular ventilation during the day, the ventilation in the cooler forest is a less efficient upward movement of air, with gas exchange restricted by reduced surface exchange area. These results, together with other recent findings, question entrenched ideas such as the thermosiphon-ventilation mechanism or the assumption that mounds function to dissipate internally produced heat. Models trying to explain the proximate mechanisms of mound building, or building elements, are discussed.