The above circumstantial evidence for the development of migration from breeding areas as birds spread from lower to higher latitudes has gained additional support in recent decades from DNA analyses. Many bird species now live in areas that remained favourable during the glaciations, as well as in higher latitude areas that could have been colonised only since glacial retreat. The majority of such species show more mitochondrial DNA variation in the refugial than in the more recently colonised areas. They are assumed to have colonised from the edge of their refugial ranges, and are like other founder populations which, because of the small number of initial colonists, show less genetic variation than the source populations. Examples include the Greenfinch Carduelis chloris in Europe (Merila et al. 1997), and the Yellow Warbler Dendroica petechia and MacGillivray's Warbler Oporornis tolmiei in Central and North America (Klein & Brown 1994, Mila et al. 2000). Reductions in genetic variance with increasing latitude, assumed to result from post-glacial leading edge spread, have also been described in plants, insects and mammals from both continents (Hewitt 2000).
Several other widespread species examined from different parts of their degla-ciated North American ranges have shown unexpectedly low levels of mitochon-drial DNA differentiation, despite, in some species, high levels of phenotypic differentiation (for Red-winged Blackbird Agelaius phoeniceus see Ball et al. 1988; for Song Sparrow Melospiza melodia see Zink & Dittman 1993a; for Swamp Sparrow Melospiza georgiana see Greenberg et al. 1998). Yet again, the implications are that large deglaciated areas were colonised by individuals from limited areas, and that such species underwent morphological differentiation only after they had spread to their current ranges, in the last 10 000 years or so. This is too short a period to expect much geographical variation in mitochondrial DNA. They contrast with some species mentioned earlier, which are likely to have spread from two or more different refuges, as they show two or more distinct types, based on mitochondrial DNA (e.g. Great Reed Warbler Acrocephalus arundinaceus in
Europe, Bensch & Hasselquist 1999; and Yellow Warbler Dendroica petechia in North America, Milot et al. 2000).
Interesting evidence for genetic isolation between northern migratory and southern resident populations of the same species has emerged for the Prairie Warbler Dendroica discolor in North America (Buerckle 1999). Within the continent, the species is resident year-round only in mangrove habitat around the Florida coastline, but migratory Prairie Warblers breed further north over a large part of the continent. Some of these migrants also winter in Florida, where they could come into contact with the resident population. The two populations are morphologically indistinguishable, but they can be separated genetically, having split from one another probably in the late Pleistocene. This situation shows clearly how a southern resident spreading northward to breed can become genetically isolated from its parent population, even though the two populations overlap in wintering range. The isolation develops presumably because the two populations breed in different regions, and at different times of year.
In some species, estimates have been made from mitchondrial DNA of both the dates and routes of spread. Thus, based on diversity patterns in mitchondrial DNA (four sections), Common Chaffinches Fringilla coelebs appear to have spread from North Africa into southern Europe within the last 370 000 years or so, and spread further north through Europe only during the past 15 000-3000 years, behind the retreating ice sheets (Marshall & Baker 1999). Only the more northern populations, breeding in the most seasonal environments, have become migratory. Similarly, studies of several North American bird species, including the Song Sparrow Melospiza melodia and various parids, have revealed basal haplotypes in Newfoundland, a known glacial refuge area, suggesting that these species may have spread from there to other parts of the continent in post-glacial times (Gill et al. 1993, Zink 1997). In these populations, spread was not only northward, but mainly westward and to some extent southward.
A similar approach was applied to two pairs of closely related plover species:
(1) the non-migratory Australasian Charadrius ruficapillus and the similar largely migratory C. alexandrinus which occurs across Australia, Eurasia and the Americas;
(2) the non-migrating C. australis of Australian deserts and C. veredus which migrates between breeding areas in eastern Asia and non-breeding areas in Indonesia and Australia (Joseph et al. 1999). With these distributions, one possibility is that the strongly migratory alexandrinus and veredus evolved from Australian ancestors through northward shifts of their breeding distributions from Australia to continental Asia. Alternatively, there may have been southward shifts in breeding distribution from eastern Asia to Australia, with the non-migratory Australian species ruficapillus and australis having arisen from migratory ancestors. In fact, DNA analyses suggest that, in both these species pairs, the more northern migratory species evolved from the non-migratory ones.
This latter type of study is as relevant to the evolution of species as to the evolution or modification of migration, for we know that changes from resident to migrant or vice versa can occur rapidly within species (Chapter 20). While the path of evolution in the plovers may well have been as DNA suggests, any of the four species could probably have become migratory if they had spread into a seasonal environment that encouraged it. In practice, many species spreading from low to high latitudes have to develop migration if they are to avoid the cold, high latitude winters. This holds in comparisons between species, as well as between different populations of the same species.
Another fact adduced in support of a tropical origin of many long-distance migrants is that long-distance tropical landbird migrants are phylogenetically more different between North America and Eurasia than are the short-distance migrants and residents of these two land masses (Bohning-Gaese et al. 1998). The longdistance migrants are likely therefore to have evolved further south, in tropical regions, where greater distances separate the two land masses, ensuring greater isolation (in evolutionary time) between their respective avifaunas. Moreover, most of the long-distance passerine migrants belong to families that are essentially tropical in distribution, such as the Parulidae and Tyrannidae in the New World and the Silviidae and Muscicapidae in the Old World. This has fuelled the view that these mainly insectivorous species have evolved from tropical families, representatives of which spread north to breed in northern latitudes (Rappole & Warner 1980).
It is in any case impossible to separate in the northern hemisphere the situation in which a tropical species spreads gradually north and develops southward migration secondarily, from that in which a species develops from within its existing range a northern migration to a new breeding area. In practice, as mentioned above, both the northward spread in breeding range and the associated southward migration were likely to develop hand in hand once the species had reached areas unfavourable for wintering.
Southern hemisphere migration systems have been less well studied, but again some of the same conclusions seem to hold, at least in South America where the question has been addressed. In this region, around 70% of southern temperate zone birds, especially insectivores, have conspecific or closely related species in the Amazonian tropics (Joseph 1997, Jahn et al. 2004).
In summary, it can be hypothesised that the majority of current bird migrations arose in the same way, with dispersing individuals spreading from lower to higher latitudes, where they could exploit the abundant seasonal resources for breeding, but returning to lower latitudes for the non-breeding season. The types of species involved depended on the diversity of source populations, the types of intervening habitats, and the types of seasonally available breeding habitats. This view of migration is centred of the notion that 'migratory genes' persist even within essentially resident populations, allowing them to continually change in migratory habits, as well as in distribution, in response to ever-changing eco-geographical conditions. It is concerned mainly with the modification of pre-existing movement patterns, and not with the evolution of the species themselves, which could, of course, have arisen at low or high latitudes, and spread from there, developing resident or migratory behaviour as the need arose.
Was this article helpful?