Let us compare a livestock farmer and a fisher. The farmer selects and breeds individuals that exhibit the most desirable characteristics. This is good practice, because it increases the prevalence of these characteristics in the next generation of the stock. In contrast, the fisher catches large, fast-growing fish, so their desirable characteristics are less likely to be passed on to the next generation of the stock (Figure 5.1). Fish that grow quickly tend to be caught sooner and therefore may produce fewer offspring. Fish that delay maturation tend to be caught before they have the chance to reproduce, so the fish that are left to breed are those that mature at a younger age. Fish that limit their current investment in reproduction in order to increase future reproductive success will often be harvested before such savings have a chance to pay dividends. The mortality imposed by fishing can therefore act as a selective force that favours slower growth, earlier maturation and higher reproductive investment.

Introduction

When John Maynard Smith (1966) wrote on sympatric speciation more than 35 years ago, he acknowledged that the argument “whether speciation can occur in a sexually reproducing species without effective geographical isolation” was an old problem and voiced his opinion that the “present distribution of species is equally consistent either with the sympatric or the allopatric theory.” Yet, from the heyday of the Modern Synthesis until relatively recently, the importance of sympatric speciation has been downplayed, and the corresponding hypotheses remained obscure well beyond Maynard Smith's seminal study.

Looking back from today's perspective, it is astounding that, for such a long period, the research community at large essentially turned a blind eye to sympatric speciation. Given the widely acknowledged difficulties involved in inferring past process from present pattern, one can only feel uneasy about a logic that claims to find evidence for the prevalence of allopatric speciation in the present-day distribution of species. To a large extent it seems to have been the scientific community's perception of the theory of sympatric speciation that has brought about a profound skepticism toward the broader empirical relevance of this speciation mode. Scientific attempts to overcome this skepticism have come and gone in waves. In the 1960s, luminaries of North American evolutionary biology pulled no punches when assessing the merit of such attempts.

A Shift in Focus

Millions of species currently exist on earth, and to secure an understanding of how all this magnificent variety arose is no small task. Biologists have long accepted Darwinian selection as the central explanation of adaptation and evolutionary change; yet, to date, no similar agreement has emerged about evolutionary processes that can create two species out of one. Almost 150 years after Darwin's seminal work On the Origin of Species (1859), conditions for and mechanisms of biological speciation are still debated vigorously.

The traditional “standard model” of speciation rests on the assumption of geographic isolation. After a population has become subdivided by external causes – like fragmentation through environmental change or colonization of a new, disconnected habitat – and after the resultant subpopulations have remained separated for sufficiently long, genetic drift and pleiotropic effects of local adaptation are supposed to lead to partial reproductive incompatibility. When the two incipient species come into secondary contact, individuals from one species cannot mate with those of the other – even if they try – or, if mating is still possible, their hybrid off spring are inferior. Further evolution of premating isolation (like assortative mate choice or seasonal isolation) and/or postmating isolation (like gametic incompatibility) eventually ensures that the two species continue to steer separate evolutionary courses.

The trigger for speciation in this standard model is geographic isolation. It is for this reason that the distinction between allopatric speciation (occurring under geographic isolation) and sympatric speciation (without geographic isolation) has taken center stage in the speciation debate.

Introduction

Extant patterns of species abundance are usually considered to be suggestive of allopatric speciation, because even closely related species are often geographically segregated (e.g., Barraclough and Vogler 2000; see Chapters 15, 16, and 17). Even though, in many cases, the ecological abutment between related species does not correspond to any obvious geographic barriers to gene flow, such patterns of geographic segregation are taken as strong indicators that speciation has occurred, either in allopatry or in parapatry. For the latter case it is assumed implicitly that there exists some sort of environmental discontinuity on either side of which different types are favored by selection or have evolved by genetic drift (Turelli et al. 2001). Even though gene flow across the environmental discontinuity can actually enhance speciation through the process of reinforcement, in these parapatric scenarios the reasons for speciation are ultimately the same as those in purely allopatric scenarios, that is, divergent evolution in different geographic regions. This has led to a common understanding that allopatric patterns of abundance between closely related species imply past events of allopatric speciation.

Rather than focusing on patterns of species abundance, recent developments in speciation theory focused on the adaptive processes and mechanisms that lead to disruptive selection and subsequent divergence of emerging new lineages in wellmixed, geographically unstructured populations. While this approach, described in Chapters 4 and 5, highlights the importance of frequency-dependent ecological interactions for evolutionary diversification, nonspatial models evidently cannot explain geographic patterns of species abundance.

Theories of speciation, in the past often couched in verbal terms, should explain how ecological divergence and genetically determined reproductive isolation evolve between lineages that originate from single, genetically homogeneous ancestral populations. As Will Provine highlights in Chapter 2, the predominant perspective for a long time was that reproductive isolation emerges as a by-product of other evolutionary processes, through the incidental accumulation of genotypic incompatibility between related species. It is easiest to imagine that such incompatibilities arise when subpopulations become geographically isolated and henceforth evolve independently: genetic distance between them is then expected to increase with time. Thus, “given enough time, speciation is an inevitable consequence of populations evolving in allopatry” (Turelli et al. 2001). On a verbal level this theory of allopatric speciation appears both simple and convincing. This apparent theoretical simplicity has contributed to the view that the allopatric mode of speciation is the prevalent one – a perspective that has found its most prominent advocate in Ernst Mayr (Chapter 2).

Unfortunately, not only is the simplicity of the usual accounts of allopatric speciation based on the poorly understood concept of genetic incompatibility, but simplicity in itself is no guarantee for ubiquitous validity. Other plausible, but theoretically more intricate, mechanisms for the evolution of reproductive isolation in the absence of geographic isolation have been proposed. Recent approaches have focused attention on adaptive processes that lead to ecological and reproductive divergence as an underlying mechanism for speciation processes – a change in emphasis that occurred concomitantly with a shift in biogeographic focus from allopatric scenarios to parapatric speciation between adjacent populations or fully sympatric speciation. This was foreshadowed by the idea of reinforcement (the evolution of prezygotic isolation through selection against hybrids) and has culminated in theories of sympatric speciation, in which the emergence and divergence of new lineages result from frequency-dependent ecological interactions.

When Terry Erwin from the Smithsonian National Museum of Natural History examined the diversity of beetles that lived on a single species of tropical trees, he found 682 different beetle species, 163 of which he classified as specialist species that lived exclusively on the particular tree species used in his study. Since there are around 50000 tropical trees species, Erwin extrapolated that there must be on the order of 7 million specialist beetle species (Erwin 1982). Using similar extrapolations, Erwin (1982) also estimated the total number of tropical arthropod species as about 30000000. While these estimates may be too high (Schilthuizen 2000; Ødegaard 2000; Novotny et al. 2002), they are mind-boggling nevertheless and serve as an illustration of the incredible amount of species diversity that exists on our planet: estimates for the total number of extant species of plants and animals range from 10 million to 100 million (May 1990; Schilthuizen 2000). It is also estimated that the number of extant species represents only about 1% of the total number of species that ever existed during the history of life on earth. Together with the common phylogenetic ancestry usually inferred for the tree of life for higher organisms, this implies that speciation must have been truly rampant during the creation and evolution of our biosphere.