Progress with the statistical analysis of primary molt

This paper considers two extensions to the statistical analysis of primary molt. The first extension relates to thedevelopment of statistical models that allow the simultaneous estimation of molt parameters for diverse groups of birds.Parameter estimation uses the method of maximum likelihood and the EM algorithm. The likelihood ratio criterion is appliedin statistical hypothesis-testing for equality of sets of parameters, and the Akaike Information Criterion is used to guide modelselection. This extension is illustrated using primary molt data for willow warblers (Pylloscopus trochilus) over 11 years. Thesecond extension is the realization that important insights into the way that primary molt is incorporated into the annual cycleof birds can be obtained by investigating the parameters of molt of individual primary feathers. This extension is illustratedusing six data sets from four species of waders, Charadriiformes.

Les G. Underhill, Lorenzo Serra, Anabela Brandao, Acta Zoologica Sinica 52(Supplement): 440-443, 2006