Bayesian inferences on the recent island colonization history by the bird Zosterops lateralis lateralis

The founding of new populations by small numbers of colonists has been considered a potentially important mechanism promoting evolutionary change in island populations. Colonizing species, such as members of the avian species complex Zosterops lateralis, have been used to support this idea. A large amount of background information on recent colonization history is available for one Zosterops subspecies, Z. lateralis lateralis, providing the opportunity to reconstruct the population dynamics of its colonization sequence. We used a Bayesian approach to combine historical and demographic information available on Z. l. lateralis with genotypic data from six microsatellite loci, and a rejection algorithm to make simultaneous inferences on the demographic parameters describing the recent colonization history of this subspecies in four southwest Pacific islands. Demographic models assuming mutation–drift equilibrium or a large number of founders were better supported than models assuming founder events for three of four recently colonized island populations. Posterior distributions of demographic parameters supported (i) a large stable effective population size of several thousands individuals with point estimates around 4000–5000; (ii) a founder event of very low intensity with a large effective number of founders around 150–200 individuals for each island in three of four islands, suggesting the colonization of those islands by one flock of large size or several flocks of average size; and (iii) a founder event of higher intensity on Norfolk Island with an effective number of founders around 20 individuals, suggesting colonization by a single flock of moderate size. Our inferences on demographic parameters, especially those on the number of founders, were relatively insensitive to the precise choice of prior distributions for microsatellite mutation processes and demographic parameters, suggesting that our analysis provides a robust description of the recent colonization history of the subspecies.

Monarchs across the Pacific: the Columbus hypothesis revisited

The 'Columbus hypothesis' suggests that the annual north-south return migration of Danaus plexippus in North America is a very recently evolved behaviour, less than 200 years old. This hypothesis rests, in part, on an analysis of the 19th century spread of the monarch across the Pacific that assumes a continuous east to west movement and is based predominantly on one publication. We review all the contemporary literature and present new analysis of the data. The movement of the monarch across the Pacific in the second half of the 19th century is best explained by a model which involves no more than three spot introductions, directly or indirectly aided by human movement, followed by natural spread of the monarch across island groups. Contemporary records refer to 'boom' and 'bust' population cycles on newly settled islands, which may have led to high rates of monarch movement. We see no evidence in the records to suggest an east to west sweep by monarch populations as suggested by the Columbus hypothesis. (C) 2004 The Linnean Society of London.