It has long been argued that populations living on the edges of the species geographical range should exhibit stronger environmentally imposed variations in abundances than populations living near the centre of this range. This so-called 'Species Range Hypothesis' is tested here for the recruitment of 62 marine fish populations of the northeast Atlantic, belonging to 17 species. The pattern of increasing recruitment variability from the centre towards the edges of the species range was modelled by a quadratic linear regression between an index of interannual variability in recruitment and an index of population position in species range. The relationship between recruitment variability and exploitation rate was also investigated. A wide range of recruitment variability was observed among the 62 populations. The expected latitudinal pattern of recruitment variability was significant when populations of the various species were pooled in the same analysis. At the species level, this pattern was found for most species, but was significant only for cod and sole. Recruitment variability was also dependent on the number of years in the time series, although this association did not perturb the relationship with population position. The analysis did not show any clear impact of fishing on recruitment variability. These results suggest that the influence of environment on the recruitment of fish populations is stronger for populations living on the edges of the species geographical range than for populations living at the centre of this range.