The present studies examined the nature of visuo-spatial working memory development using conventional visual span and spatial span measures. Children aged between 6 and 13 years, and adults aged 18-38 years were employed as participants. In Study 1, visual span, spatial span, articulation rate, and verbal fluency competencies were measured. In Study 2, visual span and spatial span maintenance was subject to five interference formats: nil, speech articulation, verbal fluency, visual masking, and spatial tapping. Distinct developmental rates were found for the two span tasks, which in the older children were correlated with the verbal fluency measure. Study 2 provided experimental evidence of the contribution of executive and spatial processes to spatial and to visual span maintenance. The results are interpreted as indicating that these memory span procedures make complex demands upon the visuo-spatial working memory architecture and consequently a precise identification of the processes that actually develop is compromised. It is suggested that a componential approach where tasks are constructed to tap specific working memory components would afford a more accurate understanding of the development of these processes.