It is clearly unwise to assume that, because some birds have altered their migrations in recent decades, all birds can do so. Most of the observed patterns of alteration involve gradual change, in which each incremental step is itself of selective advantage. Patterns that seem resistant to change are those that would involve a marked step-change in some feature, such as direction or fuel deposition, to be viable, while the intermediate steps would be lethal. The birds may thus be locked into some patterns simply because the step-changes in genetic control needed to break free of them are unlikely to arise by a single mutation. Some extreme journeys of today, involving long sea- or desert-crossings, can most plausibly be explained by gradual development, followed by eventual loss of the intermediate stages. Some of the strangest migration routes are likely to have developed in early post-glacial times, when many species were expanding into high-latitude regions recently freed of ice, but these routes might not persist today if the birds could make the step-changes in genetic control necessary to exploit more efficient options. The persistence of long, roundabout migration routes provides a cautionary reminder against the notion that birds invariably behave optimally (but see Alerstam 2001). Apparent legacies of the past are clearly evident in the migration routes of many birds worldwide. Selection cannot act on a clean slate, but must start on a pre-existing gene pool, adapted to different (historical) conditions. Much potential inherent variation may already have been lost.
Some other aspects of the spatial genetic structure of bird populations have been attributed to post-glacial colonisation and migration patterns. For example, many migratory species have more uniform population structure (often manifest in fewer subspecies) than closely related resident species (Belliure et al. 2000, Newton 2003). This may be partly because migrants intermix more, and are exposed to selection pressures in several different areas, rather than just one, but also because many migrant species have recently colonised huge areas from one or two glacial refuges. They have therefore not had the time, or consistency in local selection pressures, needed to impart phylogeographic structure to their populations (for examples from warblers, gulls and others see Helbig 2003). Second, since most post-glacial colonisation is north-south rather than east-west, phylogeographic differentiation tends to be less on the main migration axis than off it (for examples in the Yellow Warbler Dendroica petechia and North American sparrows see Helbig 2003). Third, highly migratory species have not necessarily been derived from ancestors that were themselves highly migratory, but frequently come from ancestors that were less migratory or sedentary (again, examples from DNA analyses are found among Sylvia and Phylloscopus warblers, Helbig 2003). In the evolution of migratory behaviour, therefore, the ancestry of a species (its 'phylogeny') seems of far less importance than the ecological circumstances in which it lives. On this view, many species can be treated as statistically independent units in analyses of migratory behaviour, because phylogeny does not to any measurable extent seem to constrain the evolution of migratory adaptations (Helbig 2003). The same probably holds for many other adaptations related to breeding latitude, such as duration and timing of moult, number of broods per season, and so on. In contrast, many other features of birds, such as body form or egg colour, are clearly influenced by ancestry, so that all members of a taxonomic family share the same features, and cannot be considered as statistically independent in these respects.
Many of the ideas proposed in this chapter, although based on a firm observational foundation, are necessarily speculative, but they are essential to a full consideration of the evolutionary and ecological background to current migration patterns. They show how complex migration patterns might have arisen step by step, each of which brings benefits and enhances fitness in its own right. They cannot be tested quantitatively in the field, particularly those that depend on past changes in the nature and distribution of habitats. The results are therefore less satisfying and in a sense 'less scientific' than are the quantitative or experimental studies that are possible on other aspects of migration. Yet they help us to understand how some otherwise puzzling aspects of current bird migration systems might have arisen and, if we disregard them merely because they cannot be formally tested, we risk ignoring some of the most important evolutionary aspects of the subject. Only 30 years ago, almost all aspects of the evolution and genetical control of migration were speculative, and increasingly they have been confirmed or modified by experiment or by genetic studies. Already DNA analyses of certain species have helped to define post-glacial colonisation routes, and confirm that they match some current migration routes. Perhaps, therefore, in the coming years, other as yet intractable aspects of the subject will yield to ingenuity or new methodology.
The underlying message from examination of migration routes is that small step-changes involving successive extensions, or minor modifications to existing routes, each of which brings fitness benefits in its own right, more readily occur than complex and abrupt changes. This conclusion exemplifies a more general principle of evolution, that genetic change must have a starting point, developing from the raw material already available, and within the constraints that this might provide. The idea of constraint for evolutionary change in migratory behaviour stands in some contrast to the great flexibility of those aspects of migration discussed in Chapter 20.
Was this article helpful?