I.  Experimental evidence for allopatric speciation.

    A.  Speciation in laboratory populations

        1.  Speciation as a process in nature is very difficult to study because one is never sure exactly where it is occurring; it can also take a long time.

        2.  For these reasons, some investigators have attempted to try to bring speciation into the laboratory.  In effect, they have set up experiments in which they have
                tried to cause speciation to occur, and observed what happens.  The first example I wish to discuss today is an experiment of this type.

    B.  Example:  Evolution of reproductive isolation in Drosophila pseudoobscura.  (Diane Dodd)

        1.  The rationale for the experiment is as follows:

            a.  Our general model for allopatric speciation says that reproductive isolation between two physically isolated populations will accumulate gradually as natural
                    selection causes the populations to diverge genetically.  This can be true for both postmating and premating isolation.

            b.  Dodd set up two types of Drosophila population in the laboratory.  In one type of population, flies were reared on a medium that was primarily starch based. 
                    In the other type of population, files were reared on a maltose-based medium.  Neither of these media is "normal" for flies, which in nature feed on rotting fruit. 
                    Adaptation to these new laboratory conditions was thus selected for in each type of population.

            c.  Dodd reasoned that if the populations evolved to be more adapted to the media on which they were raised, this would be similar to two physically separated
                    populations evolving in different directions.  And if in nature this should be accompanied by the development of reproductive isolation, one would expect this to
                    occur in the laboratory as well.   

            d.  Consequently, Dodd set out to determine if this in fact would occur.  In particular, she set out to determine whether adaptive divergence was accompanied by
                    the development of premating isolation.

        2.  Initial adaptive divergence.

            a.  Wild flies (D. pseudoobscura) were collected in the area of Bryce Canyon, Utah, brought to the laboratory, and were used to set up four starch-medium
                    populations and four maltose-medium populations.

            b.  These populations were maintained for a number of generations on their respective media. 

            c.  At first, population sizes were small and growth was poor, but over time, population size increased and individual growth became normal and vigorous. 
                    This result by itself indicated that adaptation to the novel media was occurring.

            d.  When the flies were grown on the medium to which they had not adapted, they grew poorly, indicating that the adaptations were unique to the medium on which
                    the populations were reared, and hence that there had been adaptive divergence between the two populations.

        3.  The next part of the experiment consisted of testing whether any reproductive isolation had built up between the populations:

            a.  Each of the starch-adapted populations was tested against each of the maltose-adapted populations ( 4 starch x 4 maltose = 16 pairwise combinations).

            b.  In any given test, flies were taken from one starch and one maltose population.

            c.  These flies were allowed to oviposit in a standard cornmeal-mollasses-agar medium, and the offspring were reared on this medium to give flies from both
                    populations a common rearing history.  This was done so that the flies from the different populations did not have different odors, resulting from being
                    grown in different media, that could serve as mating cues for discrimination.

            d.  Twelve virgin males and females from each of the two populations were introduced into a small plexiglass mating chamber.  

            e.  Flies from one population had tips of right wings clipped facilitate identification of individuals to population during observation.  Two replicates of each test
                    were performed, with each population's wings being clipped in one replicate, to make sure clipping did not influence the results (Results indicated clipping had no
                    effect on mating success.)

            f.  The flies were then observed and all matings were recorded as to whether they were homogamic (within-population) or heterogamic (between populations).

            g.  An Isolation index was then constructed:

                                I =  (homogamic matings -  heterogamic matings) / total matings.

                This index ranges from -1 to 1.  A value of 0 indicates random mating; I > 0 indicates that there is a tendency for more homogamic matings than heterogamic matings,
                i.e. that there is some degree of reproductive isolation.  Reproductive isolation is perfect when I  1.

        4.  Results

            a.  For the 16 starch-maltose population combinations, the value of I ranged from 0.18 to 0.49, with a mean of 0.33.

            b.  An appropriate statistical analysis indicated that this was a real departure from random mating and not a fluke due to small sample size.

            c.  As controls, Dodd also paired same-media populations.  These trials produced values of I ranging between 0.21 and 0.18, with a mean value of
                    approximately 0.  In other words, when comparing populations that had the same type of selection applied to them, there was no indication that
                    any degree of premating reproductive isolation had evolved.

        5.  These results indicate, then, that adaptive divergence of the maltose and starch populations was accoompanied by evolutionary divergence in mating behavior,
                    such that the tendency of maltose-adaptated individuals to accept starch-adapted individuals as mates, and vice versa, decreased. 

        6.  In other words, adaptive divergence was accompanied by the establishment of reproductive isolation, albeit not complete isolation.  For this amount of isolation to
                    develop in just a few generations, however, is quite remarkable, and clearly lends support to the notion that adaptive divergence in nature may often be accompanied
                    by the development of premating isolation.

II.  Coyne and Orr's analysis (Coyne and Orr. 1989.  Patterns of speciation in Drosophila.  Evolution 43: 362-381.)

    A.  One of the expectations under the allopatric model of speciation is that reproductive isolation (both pre- and post-zygotic) accumulates gradually as populations diverge genetically. 

        1.  Coyne and Orr examined empirically whether this is the case in the genus Drosophila. 

        2. This study represents the compilation of a tremendous amount of information generated by a large number of researchers on genetic divergence and degree of  reproductive
                isolation for 119 pairs of closely related species of Drosophila.  This type of study, of this magnitude, could only be done with Drosophila, because similar kind of information
                is not available for any other taxon.
   
    B.  Methods.

        1.  All data was taken from the literature, i.e. from previously published experiments.

        2.  For each species pair, the data used was of the following type:

            a.  Degree of genetic divergence was measured by the genetic distance between the pair, which is determined by comparing the allele frequencies of the two species at a
                    large number of electrophoretically detectable loci.

            b.  Genetic distance is 0 when the two species have identical allele frequencies at all loci, and is infinite when the two species share no alleles in common at any of the loci.

            c.  Among the Drosophila species examined in the study, the value of this index ranges from 0 to 2.

            d.  Degree of post-zygotic isolation

                i.  each species pair was crossed in two ways (reciprocal crosses): species A female x species B male, and species A male x species B female.
 
                ii.  Each of these crosses has the potential to yield males and females, and thus there are thus four sex x reciprocal cross combinations
                        (e.g. female offspring of spec. A male x spec. B female, etc.) .  

                iii.  Flies in each of these four categories were assigned a score of 1 if they were completely inviable or completely sterile, 0 otherwise.  

                iv.  These scores were then averaged to obtain a postzygotic isolation index that runs from 0 to 1.  1 means complete reproductive isolation, 0 little or none.

            e.  Degree of prezygotic isolation was measured in a manner similar to that used by Dodd: in mating cages, the numbers of homospecific and heterospecific matings
                    were recorded.  An isolation index was then constructed as:

                            I =  1 - ( number of heterospecific matings) / (number of homospecific matings)

                    This index ranges from 0 (no premating isolation), to 1 (complete premating isolation).  (Note: can have negative values for disassortative mating, but in practice this seldom occurs.)

    C.  Results 

        1.  Postmating isolation: postzygotic isolation increases gradually and steadily as a function of genetic distance.

        2.  Premating isolation: premating isolation also increases gradually and steadily as a function of genetic distance.

        3.  These results are exactly what is predicted by the theory of allopatric speciation, and hence constitutes some of the best evidence we have that much
                speciation is due to gradual accumulation of reproductive isolation between diverging populations that are separated by a physical barrier.

III.  Character Displacement and Coexistence

    A.  Character displacement and lineage splitting

    1.  When the physical barrier separating two reproductively lineages breaks down, the ranges of the two lineages will expand and overlap.
 
    2.    If the two lineages have not diverged sufficiently ecologically during separation, then they will compete where their ranges overlap.

    3.   There are two possible outcomes to competition:

        a.  extinction of one of the lineages

        b.  evolution of character displacement and resulting coexistence

    B.  The process of character displacement.
     
    C.  Example: Typhlosaurus lizards (Huey and Pianka)

        1.  Two species of legless lizzards, T. lineatus and T. gariepensis live in the Kalahari Desert of South Africa

        2.  T. lineatus has a larger geographic range.
   
        3.  The range of T. gariepensis is smaller and is contained entirely within that of T. lineatus. 

        4.  Range of T. lineatus thus consists of two regions:

            a. Area of sympatry with T. gariepensis

            b.  Area of allopatry with T. gariepensis

        5.  If character displacement has occurred, expect T. gariepensis to be more similar ecologically to T. lineatus from the area of allopatry than
                to T. lineatus from the area of sympatry.

        6.  Characteristics examined: indicative of ecological similarity and similarity of resource use

            a. Typhlosaurus lizzards feed primarily on insects; termites make up 95% of the diet

            b. Size (snout-vent length) reflects size of preferred insects and thus is an indicator of resource use

            c.  Size of insects in gut contents: captured lizzards of both species, dissected out gut contents, and measured insects in gut

        7.  Experimental procedure: sampled individuals of both species at several locations along a geographical transect.

        8.  Results

            a.  Average size of T. gariepensis more similar to that of T. lineatus from area of allopatry than that of T. lineatus from area of sympatry.

            b.  Similar pattern for size of insects in gut contents

    D.  Example: Darwin's finches (Schluter et. al. 1985)

        1.  Geospiza fuliginosa and G. fortis live in the Galapagos Islands.

        2.  G. fuliginosa (light red in figure) inhabits the island of Los Hermanos, G. fortis
             (dark red in figure) inhabits the island of Daphne, and  both species are
             found on the island of Santa Cruz.

        3.  Character examined: beak depth, which is related to the type of seeds eaten

   
        5.  This pattern indicates that the two species have undergone character divergence when they came into contact on the island of Santa Cruz.

    E.  Conclusion

        1.  These examples indicate that character displacement commonly occurs when two related species come into contact after divergence.

        2.  This pattern suggests that character displacement is necessary for the coexistence of two lineages that have recently undergone allopatric lineage splitting.

IV.  Reinforcement

    A.  The phenomenon

        1.  If a physical barrier separating two related lineages breaks down before complete reproductive isolation evolves, two outcomes are possible

            a. The two lineages may coalesce into one evolving lineage, or 

            b.  Natural selection due to reduced hybrid fitness may favor an increase in prezygotic isolation (reinforcement), and thus the completion of permanent lineage splitting.

        2.  Reinforcement is envisioned to occur in the following way:

            a.  Within each lineage, there is genetic variation for mating behavior, such that some genetic variants (discriminators) make an individual from one lineage less likely to mate with
                    an individual from the other lineage, relative to the probability of mating with an individual from the same lineage.

            b.  By definition,  non-discriminators produce more hybrid offspring.

            c.  By definition, even partial postzygotic reproductive isolation means that the fitness of hybrid offspring are reduced compared to non-hybrid offspring.

            d.  Consequently, the average fitness of discriminators will be higher than that of non-discriminators and natural selection will favor the evolution of discrimination.

        3.  Note that while natural selection may favor an increase in pre-zygotic reproductive isolation in this way, selection can not act to increase post-zygotic isolation because
                natural selection would tend to eliminate pairs of alleles at different loci causing genetic incompatibility if they cooccurred in a population.

        4.  Theoretical genetic models indicate that this process may occur, but suggest that the conditions necessary may be restrictive.

    B.  Example: Evolution of Reproductive isolation in Drosophila.  (Coyne and Orr: see above for reference)

        1.  Compared degree of reproductive isolation vs. genetic distance for pairs of taxa that were allopatric and for pairs that were sympatric.

        2.  Found that for prezygotic isolation, sympatric taxa showed greater isolation than allopatric taxa.

        3.  However, found that for post-zygotic isolation, sympatric and allopatric taxa showed no difference in isolation.

        4.  These results are consistent with the explanation that when two partially isolated Drosophila species come into contact, natural selection due to reduced hybrid
                fitness causes the evolution of increased prezygotic reproductive isolation.

        5.  Alternative hypothesis: When two species come into contact, they are more likely to coalesce into one species if the degree of reproductive isolation
                 (either pre- or post-zygotic) is small than if it is large.  Consequently, we will tend to see two separate species coexisting in sympatry only if initially they had
                diverged more rapidly than the average pair.  This bias results in an apparent greater reproductive isolation for a given genetic distance in sympatry than in allopatry.

        6.  Alternative hypothesis does not explain Coyne and Orr's results

            a.  Alternate hypothesis should apply equally to post- and pre-zygotic isolation.

            b.  Hence, expect greater isolation of both kinds in sympatry than in allopatry.

            c.  However, as the figure at left shows, only the magnitude of prezygotic isolation increases in sympatry.

            d.  Under the reinforcement hypothesis, sympatric and allopatric taxa are not expected to differ in degree of post-zygotic isolation (see A.3 above).

        7.  Conclusion:  despite theoretical predictions that the process of reinforcement is unlikely, it seems to have occurred frequently in the genus Drosophila