Gulls (Aves: Larinae) are among the best-studied of birds, yet prior a
ttempts to reconstruct gull relationships have met with little success
. In the present study I use 117 characters from the skeleton and 64 f
rom the integument to test gull monophyly and estimate gull phylogeny.
One shortest tree, requiring 9747 unweighted changes and having a CI
of 0.267, was obtained; on this tree the genus Larus is polyphyletic.
Although the tree is fully resolved, support for many of the inferred
clades is poor. In a comparison of osteological and integumentary evid
ence, I found that incongruence between the osteological and integumen
tary character sets accounts for only a minority of the total incongru
ence observed, and suggest that low between-set incongruence may be a
consequence of the low signal-to-noise ratio in each set of characters
. I also found that osteological evidence is particularly important fo
r determining higher-level structure, whereas integumentary evidence i
s important for resolving lower-level relationships within the gull gr
oup. Finally, I found that integumentary characters are not dramatical
ly more homoplasious than osteological characters, and argue that casu
al dismissal of integumentary characters as ''too labile'' is unwarran
ted. (C) 1998 The Willi Hennig Society.