Integrated data analyses are becoming increasingly common in studies of wild animal populations where two or more separate sources of data contain information about common parameters. These types of analyses provide robust parameter estimates which fully reflect all available information, as well as estimates of parameters that would be unidentifiable in a separate analysis. In this article we present an integrated Bayesian analysis of four long-term datasets (counts, two mark-recapture-recovery time series, and productivity) relating to a colony of common guillemots (/Uria aalge/) on the Isle of May, southeast Scotland. A complication when considering the dynamics of populations of this kind is the unobservable emigration of immature animals. In the analysis of mark-recapture-recovery data, the rate of emigration is frequently confounded with that of mark loss and it is only possible to estimate the product of these parameters. By combining all available data for the Isle of May guillemots in an integrated population model, we are able to estimate these parameters separately and thus obtain improved estimates of pre-recruitment emigration.