Year-to-year changes in abundance of the copepod Calanus finmarchicus in the North Atlantic are studied by means of the Continuous Plankton Recorder (CPR). Using data collected during the years 1962 to 1974, the spatial heterogeneity of long-term trends is studied by numerical analysis: Mantel test, Mantel correlogram, and spatio-temporal clustering. Results reveal that (1) interannual changes of C. finmarchicus abundance are spatially autocorrelated; (2) the spatial structures have a radius varying between 400 and 1100 km; and (3) there is a high variability in the annual changes observed between the different zones detected by clustering. These results show that observations made in the Northeast Atlantic cannot be extrapolated to the whole North Atlantic basin, and suggest that identification of the size and location of an ''homogeneous zone for long-term changes'' should be taken into account when determining factors responsible for year-to-year fluctuations in abundance of C. finmarchicus.