In the last decades, the concept of human races was considered scientifically unfounded as it was not confirmed by genetic evidence. None of the racial classifications, which strongly differ in the number of races and their composition, reflects actual genetic similarity and genealogy of human populations inferred from variability of classical markers and DNA regions. Moreover, intercontinental (“interracial”) variability was shown to be far lower than that within populations: the former constitutes 7 to 10% of the total genetic variation and the latter about 85% of it. It is believed that the low level of differentiation of regional population groups contradicts their race status and suggests a recent origin of humans from one ancestral population. The results of studies of various genetic systems are in agreement with the latter conclusion rejecting the hypothesis of regional continuity. According to this hypothesis, the populations of continents regarded as large races have developed during long evolution from local types of archaic humans, in particular, Neanderthals. Phenotypic similarity of different, sometimes unrelated, populations united into one “race” is explained by strong selection since race-diagnostic traits characterize body surface and thus are directly subjected to the influence of environmental (primarily climatic) factors. It has been recently established that variability of the most important of these traits, body and hair pigmentation, is largely controlled by one locus (MC1R), which accounts for its high evolutionary lability. Other traits used for race identification are also likely to be labile and controlled by major genes. However, the fact that the currently existing race classifications are groundless does not mean that such classifications are impossible in principle. Commonly used argumentation (races do not exist because populations are not genetically separated) does not hold water. A polytypic species is characterized by genetic continuity of allopatric populations rather than the presence of narrow genetic boundaries between them. Borderlines between races are usually conventional and arbitrary. As to intergroup variation in humans, it is indeed low but comparable with that in a number of other species. There are no obstacles to the development of genetic systematics of human races.