Darwin argued that geographic isolation (as on different islands) made speciation more likely . . . but is geographic isolation necessary for speciation?

two extremes in possibilities for formation of two sympatric species from one ancestral population . . .

• allopatric speciation (as first proposed by Darwin) -- divergence in geographical isolation followed by secondary overlap

• sympatric speciation -- splitting of a geographically continuous population into two separate species

allopatric speciation involves three steps . . .

1. separation of populations by a geographic barrier
2. evolution of ecological differences and mating differences (by natural selection and/or genetic drift)
3. secondary overlap and sympatry when geographic barriers change or after colonization

in archipelagos like the Galapagos, allopatric speciation provides a plausible explanation for the origin of species in an adaptive radiation

sympatric speciation requires two conditions . . .

1. disruptive selection for distinct ecological specializations (natural selection against individuals with intermediate phenotypes)
2. selection for mating differences (natural selection against individuals that produce intermediate offspring)

but these conditions must meet stringent requirements . . .

• during speciation, genes for ecological differences must become associated with genes for mating differences (those individuals with one kind of specialization must all have one kind of mate choice)
• but in any population (remember sympatric speciation starts with one population) genes affecting choice of mates constantly recombine with genes affecting ecological specializations
• to keep the alleles for one kind of ecological specialization and one kind of mate choice together takes extreme disruptive selection -- extreme selection against recombinants (offspring of individuals that choose mates with a different ecological specialization)
• it also takes some luck! (if by chance genes for ecological specialization and genes for choice of mates become associated with each other then speciation can evolve rapidly -- "by chance" means "as a result of genetic drift")

evidence suggests that sympatric speciation is not likely . . .

• Cocos Finch is restricted to a single isolated island (no possibility for allopatric speciation, so speciation could only occur by sympatric speciation) -- molecular differences suggest that it represents an old branch of Darwin's finches -- yet it has not split into two or more species -- it feeds on both insects and seeds -- a "jack of all trades but master of none" -- no sympatric speciation on Cocos Island

• one species of Geospiza on one island in some years has two modes of beak sizes corresponding to two kinds of seeds (a situation indicating disruptive selection) and also two song patterns -- but no reproductive isolation (instead females mate randomly with respect to beak sizes) -- disruptive selection can produce ecological differences without reproductive isolation -- hence without sympatric speciation

interim conclusion: sympatric speciation is possible but only when strict conditions are met -- natural populations of vertebrates suggest that these conditions are rarely met (if at all) -- allopatric speciation is more likely -- but it too must overcome a hurdle

the origin of new species requires the evolution of ecological differences (ecological specialization) and mating differences (reproductive isolation) -- the problem that makes speciation complicated is overcoming recombination of alleles for ecological and mating differences -- so the alleles for different ecologies are not mixed up with the alleles for different mating behavior -- recombination is the process that mixes up alleles at different loci

for a more complete picture of possibilities for speciation, we should consider a wider range of alternatives -- consider when mating differences and ecological differences might evolve in relation to allopatry (geographic separation of populations) . . .

1. both evolve in allopatry and carry over when populations become sympatric after secondary overlap (allopatric speciation in the strict sense -- discussed above)

2. only mating differences evolve in allopatry -- perhaps a result of random changes in females' preferences in small colonizing populations (genetic drift) -- following secondary overlap, reproductive isolation creates natural selection for ecological differences -- recombination is not a problem because reproductive isolation is already established (Lack's hypothesis)

3. only ecological differences evolve in allopatry -- following secondary overlap, ecological differences create natural selection for reproductive isolation -- recombination is a problem (just as in sympatric speciation in the strict sense) -- for speciation to occur, ecological differences must be large enough to produce strong disruptive selection (Bowman's hypothesis)

4. only genetic incompatabilities evolve in allopatry -- following secondary overlap, genetic incompatibilities create natural selection for mating differences -- otherwise, this case is similar to the preceding

5. ecological differences result from disruptive selection without geographical separation (allopatry) -- ecological differences then create selection for reproductive isolation -- once again, speciation can occur only if disruptive selection is strong enough to compensate for recombination (sympatric speciation in the strict sense -- discussed above)

6. evolution of ecological differences might lead directly to mating differences if the same structure or behavior is required for both (sympatric speciation without recombination)

male Darwin's finches do use beak size to identify conspecific (same species) females -- but females do not and female choice of partners probably has more influence on mating -- females use differences in the males' songs to identify species -- a recent study shows that a male's beak size influences the timing of notes in his songs -- another study suggests that male Darwin's finches learn their songs from their fathers (who would have similar beak sizes) -- female Darwin's finches prefer to mate with males like their fathers (in one case) -- so it is possible in this case that differences in beak sizes could lead directly to mating differences

a problem in generalizing this case is that no other birds are known to learn songs from their fathers (maybe one other exception) -- nor do females usually prefer songs like their fathers'

conclusion: Galapagos vertebrates so far provide no clear evidence for sympatric speciation (evolution of reproductive isolation without geographic separation -- cases 3, 4, and 5 above) -- but they do provide lots of possibilities for allopatric speciation