All of Essay Thirteen in Darwin, Dogen, and the Extremophile Choice.
… the result is not death to the unsuccessful competitor, but few or no offspring. Sexual selection is, therefore, less rigorous than natural selection. —Charles Darwin 
From the ecological standpoint, a species is a population of organisms that can’t breed beyond itself without compromising the ability of offspring to compete for resources or withstand the elements and predators they are exposed to. The genetic traits of a species are generally fine-tuned to a set of resources and conditions that, in the aggregate, define its ecological niche. Since a deviation in these traits means an organism would likely be competing in the awkward area where its niche overlaps that of another species, the natural fact that disadvantaging traits aren’t likely to show up in a next generation ensures that there’s no unnecessary overlap among well-contested ecological niches. This competitive exclusion near the boundaries of a niche establishes what’s called a resource partition, and as long as ecosystems are undisturbed, the forms and behaviours of species will remain within their defining partitions.
But of course ecosystems are not always left undisturbed; sometimes an entire species disappears, leaving resources to spare; and sometimes whole new ecological vistas open up—as maritime or mountaintop ‘islands’. When a species then finds itself largely uncontested, some divergence in its species-normal set of genetic traits can be expected. And once a critical threshold of resource-use disparity has been reached within the niche—where hybrid compromises might be a disadvantage—the splitting of its genepool, or its adaptive radiation into more efficient resource-sharing strategies, is accomplished very quickly in geological time.
If speciation events can be relatively quick, in geological time, it’s because selection, in its own way, uses ‘two hands’ for raising its partitions. In evolutionary terms, competitive exclusion implies not only that an organism’s immediate prospects must decline if its genetic endowment predisposes it to compete for resources better suited to another species, but its prospects for finding a mate will, by innate sexual preferences, become subject to selection as well. So when we now consider the critical stages of speciation, it should be kept in mind that this double jeopardy is encountered even when an organism competes across racial lines; and it is sexual selection, working directly on DNA ‘bloodlines’, which makes the definitive cut between single-species populations that are already branching due to a more immediate natural selection for the efficient use of resources.  (There will be more about sexual selection as ‘definition’ later, but even here you might see how this evolutionary dynamic is similar to the way words define the more behaviourally nuanced concepts in human cultures.)
1. Darwin, Charles. 1968 [First edition, John Murray 1859]. The Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. London: Penguin Books. p. 136.
2. The original view of speciation assumed a certain degree of geographical isolation between two populations of a single species, which allowed for character displacement, followed by renewed contact; and from here it was supposed that competitive exclusion takes over. Throughout this book I am favouring a view closer to Paul Colinvaux’s (2007, pp. 121-122.) which gives far more credit to the subtlety of co-evolution. Isolation needs not be so crudely “geographical” in very large and complex ecosystems. It is still generally assumed that novel traits favouring the exploitation of new resources, at the expense of more species-typical resources, will get quickly diluted by the interbreeding of a large population. But it was demonstrated by Patrick Bateson as early as 1982 (Gould, 1995, pp. 379-380, “The Great Seal Principle”) that sexual selection goes beyond the establishing of overt traits to ensure advantageous mating; and in fact a general rule for perhaps all sexual species is “maximal attraction to intermediate familiarity”. It’s understandable that avoidance of breeding with close family is adaptive, but why is mating with intermediate family more adaptive than mating with more distant relations, unless selection for racial disparity is at play here? Nature seems to have evolved ways to get around even “novelty dilution”, so who are we to insist on a model of speciation simply because it’s crude enough for us to represent?