This article analyses the consequences of spatial interdependence for convergence in a Solow-type growth model. In such a model a region's speed of convergence depends on its location and it can be decomposed into: (i) the speed of convergence proper, (ii) the remoteness effect, and (iii) the impact of the initial gap. Also "&sgr;"-convergence is affected by spatial interaction and we propose a decomposition to isolate the impact of spatial spillovers. Using GDP per capita of European regions, we calibrate a numerical model with parameters typically found in spatial convergence studies. We find that the remoteness effect leads to considerable variation in the speed of convergence while it marginally affects "&sgr;"-convergence. Copyright (c) 2006 the author(s). Journal compilation (c) 2006 RSAI.