Drug-resistant mutant forms of an organism are likely to be less fit than their wild-type strains in the absence of selection. Experimental work on prokaryotic organisms suggests that this is the case, but that compensatory mutations may occur which restore the fitness of mutants to that of sensitive forms. Here, we review experimental and field studies on this subject in malaria. In the rodent model Plasmodium chabaudi, a pyrimethamine-resistant mutant has been found to grow more slowly in mice than its drug-sensitive progenitor; however, following passage in the absence of the drug it grew faster, suggesting the occurrence of compensatory mutations. Similar findings were made with a chloroquine-resistant mutant. Field studies on Plasmodium falciparum have provided circumstantial evidence of a loss of fitness of chloroquine-resistant mutants, which appear to become less frequent in the parasite population following withdrawal of the drug. However, the occurrence of frequent recombination in the life-cycle of this parasite means that in natural conditions, a gene conferring resistance, once it has arisen, can then spread into a diversity of genetically distinct backgrounds which will influence its fitness and capacity to survive in the parasite population.