Theory predicts that cooperative breeding should only occur in species in w
hich certain individuals are constrained from breeding independently by som
e peculiarity of the species' ecology. Here, we use comparative methods to
examine the role of variation in ecology in explaining differences between
taxa in the frequency of cooperative breeding. We address three questions.
First, does the frequency of cooperative breeding vary at just one phylogen
etic level, or across several levels? Second, are differences in the freque
ncy of cooperative breeding among closely-related species correlated with e
cology! Last, are ecological differences between ancient lineages important
in predisposing certain lineages to cooperative breeding? We find that var
iation in the frequency of cooperative breeding occurs across all phylogene
tic levels, with 40% among families and 60% within families. Also, variatio
n in the frequency of cooperative breeding between closely related species
is associated with ecological differences. However, differences in the freq
uency of cooperative breeding among more ancient lineages are not correlate
d with differences in ecology. Together, our results suggest that cooperati
ve breeding is not due to any single factor, but is a two step-process: lif
e-history predisposition and ecological facilitation. Low annual mortality
predisposes certain lineages to cooperative breeding. Subsequently, changes
in ecology facilitate the evolution of cooperative breeding within these p
redisposed lineages. The key ecological changes appear to be sedentariness
and living in a relatively invariable and warm climate. Thus, although ecol
ogical variation is not the most important factor in predisposing lineages
to cooperative breeding, it is important in determining exactly which speci
es or populations in a predisposed lineage will adopt cooperative breeding.
Key words: birds, comparative methods, cooperative breeding, ecological co
nstraints, mating system.