Problems in the Analysis of Dispersal and a Critique on its `Heritability' in the Great Tit

(1) In data where movements between discrete points have been recorded, the set of all possible observations is finite. Studying this set of potential observations gives us insight in the limitations of actual observations. (2) In an analysis of great tit data from Wytham, Greenwood, Harvey & Perrins (1979) have demonstrated that offspring resemble their parents in the distance moved from nestbox of birth to nestbox of first breeding. They have interpreted this resemblance as heritable. It is shown here that a similar result may be obtained assuming that movements from one nestbox to another are random. That the dispersal distance of offspring is similar to that of the parents follows from the differences between the sets of all possible observations starting from the nestbox where a bird was born. Had the bird been born in another box, the set of potential observations would have been a different one. (3) The resemblance between parents and offspring in observed natal dispersal, assuming no differences in behaviour between individuals, is dependent on (i) the average distance moved, (ii) the size of the study area and (iii) the micro-distribution of nestboxes over the study area. (4) Recommendations are made as to how the problems that are outlined here can be avoided.