Infection by a parasite often induces behavioural changes in the host and these changes may benefit either the host or the parasite. However, whether these changes are active host defence mechanisms or parasitic manipulations or simply incidental byproducts of the infection is not always clear. It has been suggested that understanding the proximate mechanisms of these changes as well as comparative studies could help distinguish these alternatives better. Behavioural fever is a common response to an infection in many animals and we investigated the phenomenon in the novel host-parasite relationship between the honeybee and the temperature-sensitive microsporidian Nosema ceranae. Our results show that infected bees prefer higher temperatures and even though this seems to benefit the pathogen, the proximate mechanism underlying this change is the pathological stress underlying the infection. Especially because it is a new host-parasite relationship, it is best to label the observed behavioural change as a case of incidental benefit although this does not rule out selection acting on it. We discuss the importance of looking at the behavioural outcomes of host-parasite relationships and the importance of studying them at multiple levels for understanding their origin and maintenance.