Abstract. This paper deals with the spatial distribution pattern of the bird‐dispersed plant Daphne gnidium in a 10‐yr abandoned field under Mediterranean conditions. Colonization of Mediterranean old‐fields by bird‐dispersed plants is expected to fit a theoretical model in which (1) seed dispersal follows a negative exponential curve with the distance from the seed source and (2) seedlings are better established under perches (nucleation sensu Yarranton & Morrison 1974). However, post‐dispersal processes such as seed predation, seed germination and seedling establishment are also spatial‐dependent and can lead to spatial autocorrelation in the seedling distribution within an old‐field.Results show that both processes in the model (curve of seed dispersal and nucleation) significantly explained the spatial distribution of the seedlings, but some spatial variance remained unexplained. The semivariogram with the statistical residuals of the model detected spatial dependence at small (< 20 m) and large (> 250 m) distance intervals, indicating that some mechanisms with spatial components, apart from the curve of seed dispersal and nucleation under perches, also determined the distribution of seedlings colonizing fields. At scales below 20‐m intervals, semivariance increased indicating that similarity between plots is lost when distance increases between them. This pattern may be explained because the favourable micro‐environmental conditions for establishment produced under perches could be extended towards neighbour plots where perches were absent. A flat semi‐variogram between 20‐m and 250‐m intervals shows spatial independence along this range. From 250 m on, the semivariance increased again, indicating spatial dependence at coarse‐scale. It is possible that the colonization model failed at this scale because different spatial processes to those included in the model (perch presence and distance to the shrubland) could be controlling seed colonization at coarse‐scale.