In long term studies on birds one frequently encounters both dead reco
very data, from birds marked as young, and live recapture data, typica
lly from birds ringed as breeding adults. However the two types of dat
a have usually been analysed separately, using different computer soft
ware. In this paper we show how related but independent sets of recove
ry and recapture data may conveniently be analysed simultaneously usin
g just one computer package for model fitting by maximum likelihood. A
weighted least squares analysis provides a direct way of assessing th
e relative contributions of the two types of information. Data on the
Herring Gull Larus argentatus are used to illustrate the approach. By
showing how recovery information can be analysed by the program SURGE,
we provide a useful tool for the analysis of such data.