Daghestan contains many distinct populations characterized by different degrees of isolation, ethnic backgrounds and ecological conditions. This report introduces a large-scale family study conducted in seven Daghestan populations as well as in Moscow, using four categories of measures: (1) single gene polymorphisms, (2) anthropometric, (3) physiological, and (4) psychological quantitative traits. Such a study permits the analysis of both phenotypic and genetic variation within and between populations. The implications of such variation are discussed, especially in the context of the measures used. These derive from Soviet conceptualizations of the nervous system and related assessment of individual differences, and vary considerably from conventional Western approaches to behaviour. From the anthropometric through the physiological to the behavioural measures there was a clear trend towards increased phenotypic variation but reduced genetic variation. Possibly because the genetic variance was less on the more complex phenotypes, measures of generalized distance indicated much greater population similarity when based on phenotypic data on the genetic component of variation. Issues that arise when using behavioural traits to study human population diversity are discussed, particularly in relation to inbreeding and the specific cultural and linguistic practices in the Daghestan region.