• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Nov 28, 2006; 103(48): 18184–18189.
Published online Nov 15, 2006. doi:  10.1073/pnas.0608777103
PMCID: PMC1838727

Drosophila flies in “Evolution Canyon” as a model for incipient sympatric speciation


The genetic basis of population divergence leading to adaptive radiation and speciation is a major unresolved problem of evolutionary biology. Molecular elucidation of “speciation genes” advanced recently, yet it remains without clear identification of the gene complexes participating in reproductive isolation between natural populations, particularly, in sympatry. Genetic divergence was discovered between Drosophila melanogaster populations inhabiting ecologically contrasting, opposite slopes in “Evolution Canyon” (EC), Mt. Carmel, Israel. Interslope migration of flies is easy and verified. Nevertheless, significant interslope D. melanogaster population divergence was established at EC involving habitat choice, mate choice, thermal and drought tolerances, adaptive genes, and mobile elements. Parallel patterns of stress tolerance, habitat choice, and mate choice were demonstrated in Drosophila simulans at EC, although on a smaller scale. However, some tests for interslope genetic differentiation in Drosophila, derived from the opposite EC slopes, gave somewhat controversial results. Here we present new empirical data on interslope genetic divergence of Drosophila at EC, and summarize previous supporting and controversial results. We suggest that Drosophila populations at EC represent a rare example, demonstrating how selection overrides migration, and propose an ad hoc ecological model of incipient sympatric divergence.

Keywords: genetic divergence, incipient differentiation, natural populations, sympatric evolution

Animals' adaptation to and speciation in severely changing environments are among the most amazing phenomena in evolutionary biology. Unraveling the complex selective networks accompanying population divergence involving migration, habitat choice, sexual behavior, and stress tolerances is of primary importance in highlighting evolution in action. Especially intriguing is the problem of genetic adaptation to microsite ecological heterogeneity as a driving force in sympatric speciation. Although not long ago geographical isolation was considered a major precondition of evolution of mating isolation (13), the idea that sympatric speciation also is possible is widespread and tolerable nowadays (49). It was proposed that strong divergent natural selection might override gene flow at very small spatial scales (10, 11) and could be a primary cause of speciation (12, 13).

Several theoretical models have been proposed for evolution of premating isolation promoted by divergent ecological and sexual selections (6, 1417). However, extensive field studies on suitable experimental models are mandatory for a better understanding of this controversial scenario (8, 18, 19). Some relevant empirical studies have been published during the last two decades, but only a few dealt with incipient population differentiation in sympatry caused by ecological selection (reviewed in refs. 6 and 8). The main problems of sympatric speciation can be represented by a series of questions (8, 2022). Is sexual isolation an incidental by-product of genetic divergence independent of whether differentiation evolved by random drift or selection? Is it selected because of its contribution to higher fitness (14)? Will reproductive isolation evolve repeatedly in independent, closely related populations subjected to divergent selection? If so, will this parallel process retain reproductive compatibility between populations inhabiting similar environments (19)? Are there specific signatures that distinguish a sympatric ecological speciation from an allopatric one? What is the relative importance of ecological and sexual selection in evolution of sympatric reproductive isolation (23)? Does reinforcement (24) play a role in completion of genetic differentiation? What is the relative importance of premating and postmating isolating mechanisms (25)? The complexity of the problem follows from this rather incomplete list.

“Evolution Canyon” (EC), a microsite at Lower Nahal Oren, Israel, represents a relevant model of microscale evolution (2628). Some 2,500 species have hitherto been identified at EC in an area of 7,000 m2. Fifteen model species have been intensively studied displaying genomic, proteomic, and phenomic divergence between the adjacent slopes. The geology and macroclimate are the same, but the interslope microclimate contrast is strong due to the opposite orientation of the slopes, resulting in up to 800% higher insolation. This causes higher temperatures and drought on the south-facing slope (SFS), displaying a xeric African biota sharply contrasting with the shadier, cooler, and mesic European biota on the north-facing slope (NFS) (29). The interslope divergence in genes, genomes, populations, and species, including several Drosophila species (30), turns EC into an ecological microscale theater for studying evolution in action across life. Notwithstanding the supporting evidence published in more than 150 papers (http://evolution.haifa.ac.il), controversy is ongoing, especially in Drosophila and its potential incipient sympatric divergence in EC.

Our hypothesis was that strong microclimatic contrasts could cause interslope divergence for stress-related gene complexes, despite a small interslope distance. Substantial genetic divergence was indeed discovered between populations inhabiting the opposite slopes. Under high interslope gene flow, the evolving slope-specific gene complexes should undergo recombinational collapse unless ecological selection overrides gene flow. Therefore, testing for divergence of adaptive traits seems to be more relevant evidence for interslope selection than scoring genetic distances estimated by using neutral molecular markers. This consideration may explain why tests for interslope divergence showed relatively high heterogeneity among molecular markers, whereas adaptive traits and relevant candidate genes showed quite consistent patterns (among collection years, traits, and, to some extent, tested candidate genes). Such results fit the expectation that, under sympatry, the selection-driven genetic divergence between populations (species) should display higher across-genome heterogeneity than that under allopatry (12, 13).

Here we present new evidence on genetic divergence at EC and highlight the accumulated evidence (both supporting and discrepant) on Drosophila obtained at the Institute of Evolution (Haifa, Israel) and other laboratories worldwide. We present a clarifying model of Drosophila incipient sympatric divergence in EC and explore whether the discovered differentiation is ephemeral or lasting over years.

Results and Discussion

Adaptation to Stressful Environments.


Temperature and humidity are the most important environmental factors affecting adaptive strategies of such a small insect as Drosophila. Therefore, Drosophila adaptation to temperature and drought stress is the target system for adaptive trait complex examination. The abundance of Drosophila at EC manifested a marked seasonal variation associated with temperature and humidity (30). To evaluate how long-term treatment by elevated temperature applied during the entire life span affects longevity, the effect of high temperature on life span was scored in SFS and NFS EC Drosophila lines in three consequent generations. Although NFS flies (males but not females) showed higher longevity at the control temperature, the relative change in longevity at 29°C as compared with the control proved higher in SFS flies (Table 1). The increase in thermotolerance of SFS flies may derive from physiological modifications and/or from short-term response to selection, presumably higher in SFS flies due to their higher genetic heterogeneity.

Table 1.
Change in life span in D. melanogaster lines caused by lifetime maintenance at an elevated temperature

The first results evaluating the reaction of D. melanogaster from EC to heat treatments were published in 1998 (31). Significant interslope divergence for stress-tolerance and adaptive behavioral traits has been demonstrated. Fitness traits included viability and longevity changes caused by short-term and lifetime temperature treatments. The revealed lower influence of heat treatment on the longevity of the SFS rather than on the NFS flies corroborates well with the conclusions reached in artificial selection experiments: Selection for stress resistance may increase longevity (32).

Extensive experiments tested whether the revealed interslope divergence with respect to thermotolerance is lasting over the years. These tests included flies sampled from EC in 1997–2004. The outcomes were in line with other studies in many respects: an inverse correlation between survival and heat-shock temperature, male–female differences in thermotolerance, and inducible thermotolerance (33). Slope, sex, year of collection, duration of heat shock, and pretreatment all affected the survival. In the vast majority of the SFS–NFS comparisons across years, SFS flies consistently exceeded NFS flies in basal and inducible thermotolerance after diverse heat shocks, no matter whether isofemale lines (34), synthetic populations, or inbred lines (33) were analyzed. The results on fresh and laboratory stocks suggest that interslope difference in thermotolerance is genetically based.

Drought and starvation resistance.

Drosophila adults are very sensitive to desiccation, and survival in a dry and hot environment determines their geographic and microhabitat distribution (35, 36). In our studies, resistance to desiccation was estimated by comparing the median lethal time of the EC D. melanogaster lines subjected to starvation at normal temperature. Tolerance to desiccation plus starvation treatment was significantly higher in SFS lines (Fig. 1). Rapid loss of stress tolerance (to desiccation and starvation stresses) is known to occur during adaptation of D. melanogaster to laboratory conditions (37). Thus, the fact that in our tests a 1.5-year maintenance of flies under standard laboratory conditions has not abolished the differences in reaction to drought treatments means that the observed dissimilarities reflect the specificity of genetic adaptations to the strong interslope microclimatic contrasts.

Fig. 1.
Distribution of D. melanogaster lines for tolerance to drought stress (1998 collection). The light columns represent SFS, and the dark columns represent NFS. The “slope” effect was significant at P = 10−6.

Pattern of desiccation resistance in EC Drosophila revealed in our previous tests (31) was quite complicated: NFS lines manifested higher survivorship than the SFS lines when subjected to starvation at conditions in which only one factor of the two, either temperature or humidity, was limiting. SFS lines manifested a much higher stability when the reaction to desiccation was measured under elevated temperature or the reaction to heat treatment was measured under drought stress.

Fitness in fecundity, fertility, and development.

The effect of temperature was also scored by using other fitness-related traits. A considerably lower fecundity was displayed by NFS flies maintained at normal temperature, but no interslope difference in fertility was found. Significant difference in larval viability, observed when the development was at the normal temperature (25°C), disappeared at the elevated temperature (29°C). This pattern is evidently caused by the higher sensitivity of NFS larvae to increased temperature. SFS flies develop significantly slower at both normal and elevated temperatures; the interslope difference increases at elevated temperatures. Significant interslope differences were also found for the variance in developmental time: at both temperatures, variation among SFS flies was higher than that of the NFS. The prolongation of the developmental period and its inter-individual variation may display an adaptive strategy, reducing the chance that the entire progeny of a female will be eliminated due to a sudden severe stress at a critical period of development. This trait complex (increased fecundity, thermotolerance, prolongated pupation, and a higher between-individual variation in the time of development of SFS flies) suggests a more secure survival of flies inhabiting the more climatically variable and stressful SFS (38, 39).

Interslope migration.

Significant interslope divergence for a complex of fitness-related traits led to the hypothesis that strong differential selection for stress tolerance could also promote microevolution of behavioral traits facilitating the differentiation in statu nascendi: (i) reduced migration rate, (ii) habitat choice, and (iii) positive assortative mating (Fig. 2). The interpretation of the revealed interslope divergence strongly depends on whether migration is indeed common between the slopes, or by contrast, if differential selection has resulted in a genetically determined reduction in the migration rate as predicted by some theoretical models (40).

Fig. 2.
Potential factors involved in population divergence in the EC.

To evaluate the migration rate, T. Pavlicek and collegues (41) used the capture-mark-release-recapture method. The conclusion was that D. melanogaster + Drosophila simulans flies from EC display substantial interslope migration. The proportion of migrants from the SFS to the NFS was estimated as 9–10%, and from the NFS to the SFS, 1–1.5%. In laboratory experiments we also examined the migratory activity of EC flies and flies collected from an open forest park population. No differences in migratory activity were revealed (42).

These high estimates of migration rates, taken together with the results on interslope adaptive divergence, suggest that strong microclimatic natural selection overrides (presumably with other behavioral mechanisms) the effect of migration.

Habitat choice.

Notwithstanding the high interslope migration, the SFS and NFS populations at EC are not panmictic. In a large-scale experiment (31), we found that adaptation to the opposite EC slope conditions has presumably resulted in divergence for habitat preferences: a significantly higher mean temperature was preferred by SFS than by NFS females with a general increase in the preferred laying temperatures with the altitude. The latter trend was much stronger for the SFS than for the NFS, reflecting the “slope × altitude” interaction. Significantly higher between-line variation was found on the SFS. A parallel result was displayed by D. simulans: significantly higher temperatures were preferred by the SFS than by NFS females. The revealed differentiation in the preferred oviposition temperature of flies raised over 1.5 years in standard conditions indicates the existence of a genetic basis of behavioral adaptation to the microclimatic slope-specific environments.

Sexual behavior and assortative mating.

The most exciting findings on EC flies come from sexual behavior experiments indicating that the EC population cannot be considered panmictic. These findings include positive assortative mating (22, 43), interslope differences in mating propensity, sexual discrimination and reproductive behavior (39), and slope specificity in the courtship song pattern (K.I., unpublished data). Some discrepant and controversial data on nonrandom mating in EC flies were also obtained (see below) inspiring further investigations.

Mating preferences.

Standard schemes of multiple-choice and single-choice experiments were used to explore interslope sexual isolation. The tests conducted with slope-specific synthetic populations detected a significant preference of sexual partners originating from the same slope (22). Results of single-choice tests corroborated those of multiple-choice tests. However, no significant departure from random mating was found in mixtures of flies derived from the opposite EC slopes in 1995 in a study by T. Panhuis et al. (44). Some peculiarities in the procedure that could affect the results should be noted. In our tests (22, 39, 43, 45), flies were observed until ≈50–60% of the maximal number of couples were formed to avoid a nonchoice situation, unlike the protocol of T. Panhuis et al. (44) where mating pairs were aspirated until a maximum of 50 pairs were collected (more than 78% of the maximal couples number), or 1 hr had passed from the beginning of test. Recent experiments for premating isolation in flies from EC, conducted in two laboratories (Simon Fraser University, Burnaby, Canada, and University of Haifa, Haifa, Israel), under a variety of protocols, detected significant assortative mating in three of six single choice tests in Burnaby, suggesting that the populations are behaviorally differentiated in some manner (45). Four of six single choice trials using SFS females produced a significant proportion of homotypic mating. NFS females also exhibited a significant excess of homotypic mating in four of six trials. However, this study failed to detect assortative mating in double choice tests.

Further experiments were undertaken with fresh collections of D. melanogaster and D. simulans (43). Significant deviation from random mating was found in both female-choice and male-choice tests in the majority of crosses of lines from both slopes and in both spring and fall collections. The predominating pattern was a significant excess of pairs formed by partners of the same slope. In total, of 200 tests with lines from both seasons and slopes, 152 showed significant positive assortative mating, 17 showed significant negative assortative mating, and 31 displayed random mating. In female-choice tests, higher between-line heterogeneity was found for SFS flies than for NFS flies. The significance of other components of variation and their interaction was tested by using log-linear analysis with the following independent components of variation: choosing sex, slope, season, choosing line, and partner line. It appeared that, in NFS flies, all considered factors and interactions were significant, whereas in SFS flies the mating preference and its uniformity tended to increase from spring to fall.

The multiple-choice test in D. simulans revealed a significant excess of pairs formed by partners originating from the same slope. These results correspond to the previous findings (22, 39) on D. melanogaster and allow further generalization that differential microclimatic selection can promote evolutionary changes fostering partial premating isolation.

Anatomy of sexual behavior and courtship song patterns.

Findings on mate choice in EC flies call for analysis of the behavioral “anatomy” of this nonrandom mating. We studied sexual and reproductive behaviors in a nonchoice situation in D. melanogaster from EC (39) for all possible mating combinations.

Sexual behavior.

We estimated male mating propensity based on the courtship latency and duration of copulation. Males from the SFS showed high sexual activity in both of these measures. The females' receptivity was determined by measuring the mating speed and courtship duration. SFS females manifested higher receptivity with SFS males than with those of the NFS, whereas NFS females did not show differences in receptivity toward both types of males. Seemingly, SFS females received more stimulation from homotypic SFS males than those of the NFS. Indeed, SFS males spent more time courting SFS females. The lowest mating success among the four mating combinations was observed in the NFS female × SFS male combination. Noteworthy, homotypic SFS pairs mated much faster than other combinations. These results definitely reflect the effect of female origin and dependence of SFS female reaction on male origin.

Among the rejection elements of female behavior, decamping had a significant effect for NFS females rejecting the courtship of SFS males. Decamping is actually an active avoidance that may lead to the loss of the sexual partners' contact in nature. Thus, peculiarities of female courtship behavior and data at the time of pair formation suggest that SFS females are more receptive and less discriminating than NFS females.

The courtship song in Drosophila plays an important role in mating success. The signal produced by wing vibration during courtship consists of pulse and sine components. These components are considered factors affecting females' receptivity (46). Especially important is the length of the interpulse interval (IPI), which is measured as the distance between two consecutive peaks of pulses. Significant interslope differences were found in male courtship song parameters: differences in the mean IPI between SFS and NFS males courting females of their own origin, and decreases in IPI demonstrated by SFS males exposed to females from the opposite slope (K.I., unpublished data). SFS males displayed plasticity (capability to change IPI) of the pulse song in such a manner that it becomes different from their typical signal but remarkably close to the IPI of NFS males. Such plasticity is especially important in light of the view that IPI is a species-specific trait displaying low phenotypic and additive genetic variability (47). The results on the male's courtship song parameters characterizing the intensity of stimulation correspond to previous behavioral findings (39), in which SFS males displayed a higher mating propensity. Differences in courtship song features might be associated with females' discrimination behavior.

Reproductive behavior.

We hypothesized that specific patterns of sexual behavior are closely associated with different adaptation strategies of flies from the opposite EC slopes. The elucidation of the relationships among dynamics of egg laying, fecundity, and repeated mating indicated that neither female or male origin, nor their interaction, had any effect on fecundity; females from same-slope pairs manifested the most divergent scores. SFS females characterized by increased egg laying speed also displayed the shortest remating time. The average time of remating for NFS females was twice that of SFS females, and the origin of males was not important in this case. The analysis of sexual and reproductive behavior in a nonchoice situation indicates that mate choice in EC flies may depend on differences in mating propensity and discrimination, which could contribute as well to asymmetric pattern of sexual isolation (39).

Genetic Differentiation.

Interslope variation at microsatellite loci.

Genetic changes associated with the local adaptation of D. melanogaster in EC were examined by using microsatellite markers (48). A strong interslope divergence was found in this test, but the two populations did not differ in gene diversity, heterozygosity, variance in the maximal repeat number, or number of alleles. Microsatellite variability depended on chromosome location, and the number of alleles was correlated with the recombination rate per DNA unit length. I. Colson (49) tested the genetic differentiation at microsatellite loci in D. simulans and D. melanogaster originating from EC. In D. melanogaster, sampled in 1997, significant differences in allele frequencies between the slopes were found in four loci, but in the 1998 collection, these differences were not revealed. Two microsatellite loci showed significant interslope differentiation in D. simulans, but the general conclusion was that the frequencies of the same alleles were enriched in the two slopes. Genetic differentiation between D. melanogaster and D. simulans populations from both slopes were tested also by C. Schlötterer and M. Agis (50) using 48 microsatellite markers; interslope differentiation was reported as low.

Interslope variation of candidate genes.


Interslope divergence in the regulatory region of hsp70Ba, which encodes for the major inducible heat-shock protein of Drosophila, was found in D. melanogaster from EC (48). Multiple instances in which disruption of hsp70 regulatory regions underlies natural variation in expression of Hsp70 were previously described. Populations in EC proved polymorphic for a 1.2 kb P element in the hsp70Ba promoter. Flies from the cooler NFS accumulate 22% less Hsp70 protein than those from warmer SFS. P insertions reduced Hsp70 levels, which may be beneficial in the absence of repeated extreme stress on the NFS.


T. Panhuis et al. (44) looked for interslope divergence in a group of the accessory gland proteins, Acps, known to rapidly evolve in Drosophila. Among different Acps, several have been identified to affect female remating receptivity, egg laying rate, longevity, sperm storage and utilization, and sperm defense during sperm competition. It was expected that interslope Acp divergence would reflect the extent of reproductive isolation. However, no evidence of genetic differentiation between NFS and SFS populations (sampled in 1995 and maintained under laboratory conditions for 5 years) was found in this gene family.


Candidate genes presumably contributing to genetic variation in sexual behavior were chosen for molecular analysis: period and desaturase (51). These genes are known to include polymorphic repeated sequences, insertions/deletions, and nucleotide substitutions. Period is sex-linked, affects the courtship song in a species-specific manner, and encodes for a transcription cofactor involved in circadian rhythm (52). Desat2, responsible for the female cuticular hydrocarbon synthesis, plays an important role in male mate choice (53). The idea was that their polymorphism might affect behavioral peculiarities in flies living on opposite slopes.

Indeed, interslope differences in the period sequence encoding the (Thr-Gly)n repeat (exon 5) were established. Two variants of this repeat, n = 20 and n = 17, which are abundant in natural populations of D. melanogaster in Europe and northern Africa (54), were found to predominate in EC. The less abundant “European” variant, n = 20, appeared among NFS flies nearly threefold more compared with the SFS, presumably reflecting some advantages of the n = 20 to flies inhabiting the NFS. Female-choice tests showed that NFS females distinguish between males with specific per alleles (n = 17 vs. n = 20) as well as between males originating from opposite slopes. Females from the SFS were less discriminating and did not manifest deviation from random mating. The ability of NFS females to distinguish between different per alleles may be one of the mechanisms, providing needed allele balance to ensure an appropriate adaptability to the local conditions. The mechanism of possible adaptive effects of (Thr-Gly)n size variation may correlate with flexible conformations of poly(Thr-Gly) peptides at higher temperatures.


Testing the promoter- and exon1-containing parts of the desat2 gene revealed size polymorphism of PCR products at EC (collection 2001) but only among SFS flies. The same 16-bp deletion at the same position of the promoter as the one revealed by S. Fang et al. (55) was found on both slopes. This deletion is known to cause differences in female cuticular hydrocarbon synthesis. Another deletion (34-bp length) leading to the appearance of stop-codon exon1 was found in a smaller band, specific to SFS flies (I. Zamorzaeva, personal communication).

Expression of heat shock candidate genes.

Tests for association between the expression of Hsp genes and thermotolerance in Drosophila collected in EC during 1997–2000 were executed by using Northern blot analysis (I. Baca and J. Carmel, personal communication). In a joint scoring of Hsp gene expression and acquired thermotolerance (flies collected in 2000), higher expression displayed by SFS flies was positively correlated with survivorship. Enhanced expression of Hsp genes is regulated largely by the heat shock transcription factor (Hsf) (56). In our tests, the Hsf expression level was higher for SFS flies and correlated with survival rates in the compared populations under heat treatments. Genotypic differences were found in the dynamics of Hsp83 expression after moderate, severe, or combined heat shock treatments. Expression of the Hsp83 gene was higher in NFS flies at 36°C but decreased under more severe subsequent treatment. In contrast, the expression of the Hsp83 gene in SFS flies greatly increased when exposed to severe heat shock. Expression of Hsp genes Hsr-ω, Hsp23, and Hsp40 was examined in lines from the 2004 collection after a treatment of 36°C applied for 1 hr. Basal and acquired thermotolerance scores showed high positive correlation with induced expression of Hsp40. Mean thermotolerance and expression for the SFS flies exceeded significantly those for the NFS samples.

The Interplay Between the Different Factors: Migration, Selection, and Preservation of the Ongoing Differentiation.

The main goal of this research program (more than 30 publications from different laboratories since 1996) is to characterize the Drosophila natural population structure in EC. Our hypothesis was that strong interslope microclimatic contrasts should cause differential selection for stress-related gene complexes, resulting in interslope genetic divergence for stress tolerance, accompanied by behavioral divergence and coupled with incipient sexual isolation (see Fig. 2). Substantial genetic divergence was indeed discovered between populations inhabiting the opposite slopes. Numerous differences in stress-related traits were found in our comparisons. Extensive literature on thermal adaptation, including behavioral aspects, was reviewed by A. Hoffmann et al. (57). Population differentiation along ecological/thermal gradients in Drosophila was established by V. Loeschcke and A. Hoffmann groups (58, 59). In some cases (60), the distance scale was just a few kilometers.

Under high interslope gene flow, the evolving slope-specific adaptive gene complexes should undergo recombinational collapse even in the presence of strong differential selection. We assumed that preservation of the observed adaptive divergence was facilitated by the evolved deviation from panmixia and random dispersal. In this scenario, adaptive divergence can withstand the destructive effects of migration and recombination. Our model to explain the accumulated evidence uses a combination of strong ecological selection and a complex mechanism of restricted gene flow between the slopes, which includes habitat choice and nonrandom mating (Fig. 2). Between-slope asymmetry established in sexual and reproductive behavior probably reflects differences in adaptive life strategy for flies living on the ecologically contrasting slopes. Increased fecundity, short remating time, higher receptivity, and reduced discrimination, peculiar to the SFS females, on the one hand, and higher sexual activity of SFS males, on the other hand, should contribute to higher fitness as compensatory mechanisms providing population survivorship in stressful habitats. In addition, strong discriminative behavior of NFS females toward SFS males, disruptions in important parameters of the courtship song revealed in SFS males, and asymmetry in migration rate may serve as primary factors in the formation of incipient premating isolation.

The postulated ecological selection driving interslope divergence for adaptively important gene complexes should not necessarily be accompanied by differentiation of selectively neutral markers, unless the latter are in linkage disequilibrium with loci subjected to differential selection. This last condition can also persist despite migration, but only under very tight linkage and strong selection. Indeed, in a number of Drosophila genes, linkage disequilibria were found to decay within a few kilobases or even 1 kb (61). Therefore, testing for divergence of adaptive trait complexes should provide much more relevant evidence for interslope selection than scoring genetic distances based on neutral molecular markers. Sympatric species, unlike allopatric ones, usually show localized interspecific genomic differentiation (62). This seems to explain why tests for interslope divergence in EC were of relatively low reproducibility when neutral markers were used, whereas adaptive traits and relevant candidate genes showed quite consistent patterns among species, collection years, seasons, traits, and, to a certain extent, tested genes.

The accumulated evidence and the proposed model indicate that partial sexual isolation between the populations from the opposite slopes is an important player in preserving the evolving adaptive complexes from recombinational collapse. A record of observations for Drosophila presented here, and similar evidence from other organisms in EC (29), suggests that incipient sympatric divergence is ongoing in the canyon. However, we cannot exclude that the observed differentiation reflects a dynamic balance between several factors (i.e., migration, selection, recombination, habitat choice, and partial isolation) rather than ongoing speciation. The revealed interslope divergence calls for further critical experiments to validate the obtained pattern on other Drosophila species and populations from other canyons, and to compare competitive explanatory scenarios.


Despite dozens of years of speciation studies, several questions remain largely unanswered for both speciation (17) and adaptation in general. First, does divergence typically proceed by the substitution of many genes having a minor effect, or can genes with large effects contribute to adaptation? A recent theory on the optimum phenotype suggests that adaptation will cause an exponential distribution of gene effects, with many factors having small effects and a few having large effects (63). Second, are changes involved in speciation simply an extension of the adaptive divergence of populations? Analysis of sterility and inviability seems to reveal a distinction between interspecific and intraspecific differences (64). Third, what are the roles of linkage, dominance, and epistasis in adaptive speciation? Are they incidental properties of newly evolved genes, or are they directly selected? There may be a tight linkage between functionally related loci because of recent tandem duplication, or it may have an adaptive explanation because clustering of loci into linked blocks will facilitate the build-up of correlations between co-adapted alleles by selection in the face of gene flow (65, 66). Epistasis is particularly fundamental to speciation (a product of maladaptive gene combinations in hybrids that causes reproductive isolation) (67). The divergence in adaptive traits within a population at a microsite makes the EC system a promising natural model to bridge theory and evidence in an in-depth analysis of adaptation and speciation under heterogeneous stressful conditions. The genomic scan of the adjacent Drosophila populations at EC should help identify the loci pertaining to divergence or/and sexual isolation (28). Further experiments will genetically dissect the slope-specific divergence for stress-resistance traits and behavioral adaptations, including habitat choice and mate choice, causing premating isolation.

Materials and Methods

Reaction to heat treatment has been compared between D. melanogaster isofemale lines originated from the opposite slopes of EC. The experimental design was based on previous Drosophila thermotolerance studies (68). Longevity at elevated temperature applied during the entire life span starting from the embryo stage was scored in three consequent generations on males and females. Flies (six NFS and seven SFS lines) were kept in two regimes (25°C for control and 29°C for treating groups) until death; vials with media were renewed thrice a week; the dead flies were recorded every day. The scores were conducted by using 50 replicates per line with a single pair (male + female) in a separate vial.

Tolerance to progressive desiccation stress was tested in EC D. melanogaster (20 lines from each slope) by using chronic stress, when adults (males and females separately) are exposed to desiccation without food (69, 70). Resistance to desiccation was estimated by comparing the median lethal time of the flies subjected to starvation at normal (25°C) temperature. Vials with 20 flies of the same sex and age were covered with a single layer of cheesecloth, and inverted on a wire platform inside desiccators with fresh Silicagel blue (3–6 mm; 700 g; relative humidity, 0.1–0.2%). The relative humidity measuring instrument was the Testo 610 (Testo, Lenzkirch, Germany). Control vials were placed in desiccators containing distilled water (relative humidity, 99.9%). Adult mortality was scored at 1-hr intervals, and the time taken for half the flies to die was linearly interpolated for each vial.


We thank V. Loeschcke, J. David, and P. A. Parsons for important comments on the manuscript and M. Noor, R. A. Krebs, and C. P. Kyriacou for constructive criticism and helpful suggestions. This work was supported by Israel Science Foundation Grant 601/03-17.3, United States–Israel Binational Science Foundation Grant 9800443, the Ancell–Teicher Research Foundation for Genetics and Molecular Evolution, and the Israeli Ministry of Absorption.


Evolution Canyon
north-facing slope
south-facing slope.


The authors declare no conflict of interest.


1. Mayr E. Animal Species and Evolution. Cambridge, MA: Harvard Univ Press; 1963.
2. Rice R, Hostert E. Evolution (Lawrence, Kans) 1993;47:1637–1653.
3. Coyne JA. Evolution (Lawrence, Kans) 1994;48:19–30.
4. Jiggins CD, Mallet J. Trends Ecol Evol. 2000;15:250–255. [PubMed]
5. Turelli M, Barton NH, Coyne JA. Trends Ecol Evol. 2001;16:330–343. [PubMed]
6. Via S. Trends Ecol Evol. 2001;16:381–390. [PubMed]
7. Mallet J. J Evol Biol. 2001;14:887–888.
8. Schluter D. Trends Ecol Evol. 2001;16:372–380. [PubMed]
9. Gavrilets S. Evolution (Lawrence, Kans) 2003;57:2197–2215. [PubMed]
10. Bush GL. Ann Rev Ecol Syst. 1975;6:339–364.
11. Endler JA. Geographic Variation, Speciation, and Clines. Princeton: Princeton Univ Press; 1977. [PubMed]
12. Barton NH, Charlesworth B. Ann Rev Ecol Syst. 1984;15:133–164.
13. Schilthuizen M. BioEssays. 2000;22:1134–1141. [PubMed]
14. Kondrashov A, Kondrashov F. Nature. 1999;400:351–354. [PubMed]
15. Gavrilets S, Waxman D. Proc Natl Acad Sci USA. 2002;99:10533–10538. [PMC free article] [PubMed]
16. Doebeli M, Dieckmann U. Nature. 2003;421:259–264. [PubMed]
17. Coyne J, Orr HA. Speciation. Sunderland, MA: Sinauer; 2004.
18. Tregenza T, Butlin RK. Nature. 1999;400:311–312. [PubMed]
19. Rundle HD, Nagel L, Boughman JW, Schluter D. Science. 2000;287:306–308. [PubMed]
20. Dobzhansky T. Genetics and the Evolutionary Process. New York: Columbia Univ Press; 1970.
21. Noor MAF. Heredity. 1999;83:503–508. [PubMed]
22. Korol A, Rashkovetsky E, Iliadi K, Michalak P, Ronin Y, Nevo E. Proc Natl Acad Sci USA. 2000;97:12637–12642. [PMC free article] [PubMed]
23. Boughman JW. Nature. 2001;411:944–948. [PubMed]
24. Dobzhansky T. Genetics and the Origin of Species. New York: Columbia Univ Press; 1937.
25. Arnquist G, Edvardsson U, Friberg M, Nilsson T. Proc Natl Acad Sci USA. 2000;97:10460–10464. [PMC free article] [PubMed]
26. Nevo E. Proc R Soc London Ser B; 1995. pp. 149–155.
27. Nevo E. Theor Pop Biol. 1997;52:231–243. [PubMed]
28. Nevo E. Proc Natl Acad Sci USA. 2001;98:6233–6240. [PMC free article] [PubMed]
29. Pavlicek T, Sharon D, Kravchenko V, Saaroni H, Nevo E. Isr J Earth Sci. 2003;52:1–9.
30. Harry M, Rashkovetsky E, Pavlicek T, Baker S, Derzhavets E, Capy P, Cariou M-L, Lachaise D, Asada N, Nevo E. Biologia. 1999;54:683–703.
31. Nevo E, Rashkovetsky E, Pavliceck T, Korol A. Heredity. 1998;80:9–16. [PubMed]
32. Parsons PA. Behav Genet. 1993;23:231–238. [PubMed]
33. Rashkovetsky E, Iliadi K, Michalak P, Lupu A, Nevo E, Feder ME, Korol A. Heredity. 2006;96:353–359. [PubMed]
34. Lupu A, Pechkovskaya A, Rashkovetsky E, Nevo E, Korol A. Mutagenesis. 2004;19:383–390. [PubMed]
35. Hoffmann AA, Parsons PA. J Evol Biol. 1993;6:643–657.
36. Parsons PA. Ann Rev Ecol Syst. 1983;14:35–55.
37. Hoffmann AA, Hallas R, Sinclair C, Partrige L. Evolution (Lawrence, Kans) 2001;55:436–438. [PubMed]
38. Rashkovetsky E, Iliadi K, Nevo E, Korol A. Dros Inf Serv. 2000;83:138–140.
39. Iliadi K, Iliadi N, Rashkovetsky E, Minkov I, Nevo E, Korol A. Proc R Soc London Ser B; 2001. pp. 2365–2374. [PMC free article] [PubMed]
40. Wiener P, Feldman MW. Evol Ecol. 1993;7:251–269.
41. Coyne JA, Milstead B. Am Nat. 1987;130:70–82.
42. Iliadi K, Iliadi N, Rashkovetsky E, Girin S, Nevo E, Korol A. Behav Genet. 2002;32:173–180. [PubMed]
43. Singh SR, Rashkovetsky E, Iliadi K, Nevo E, Korol A. Behav Genet. 2005;35:753–764. [PubMed]
44. Panhuis TM, Swanson WJ, Nunney L. Evolution (Lawrence, Kans) 2003;57:2785–2791. [PubMed]
45. Drake AE, Rashkovetsky E, Wong D, Rundle HD, Mooers AO. J Evol Biol. 2005;18:1123–1129. [PubMed]
46. Kyriacou CP, Hall JC. Nature. 1984;308:56–65.
47. Ewing AW. Biol Rev. 1983;58:275–292.
48. Michalak P, Minkov I, Helin A, Lerman DN, Bettencourt B, Feder ME, Korol A, Nevo E. Proc Natl Acad Sci USA. 2001;98:13195–13200. [PMC free article] [PubMed]
49. Colson I. Mol Ecol. 2002;11:1311–1316. [PubMed]
50. Schlötterer C, Agis M. Mol Biol Evol. 2002;19:563–568. [PubMed]
51. Zamorzaeva I, Rashkovetsky E, Nevo E, Korol A. Mol Ecol. 2005;14:3235–3245. [PubMed]
52. Gleason JM, Powell JR. Mol Biol Evol. 1997;14:741–753. [PubMed]
53. Coyne J, Crittenden AP, Mah K. Science. 1994;265:1461–1464. [PubMed]
54. Sawyer LA, Hennessy JM, Peixoto AA, Rosato E, Parkinson H, Costa R, Kyriacou CP. Science. 1997;278:2117–2120. [PubMed]
55. Fang S, Takahashi A, Wu C-I. Genetics. 2002;162:781–784. [PMC free article] [PubMed]
56. Pirkkala L, Nykanen P, Sistonen L. FASEB J. 2001;15:1118–1131. [PubMed]
57. Hoffmann AA, Sorensen JG, Loeschcke V. J Therm Biol. 2003;28:175–216.
58. Hoffmann AA, Shirriffs J, Scott M. Funct Ecol. 2005;19:222–227.
59. Sarup P, Sorensen JG, Dimitrov K, Barker JSF, Loeschcke V. Heredity. 2006;96:479–486. [PubMed]
60. Bubliy OA, Loeschcke V. Biol J Linn Soc. 2005;84:119–136.
61. Langley CH, Lazzaro BP, Phillips W, Heikkinen E, Braverman JM. Genetics. 2000;156:1837–1852. [PMC free article] [PubMed]
62. Drès M, Mallet J. Philos Trans R Soc London B. 2002;357:471–492. [PMC free article] [PubMed]
63. Marin I. J Theor Biol. 1997;188:515–524. [PubMed]
64. Wu C-I. J Evol Biol. 2001;14:851–865.
65. Korol AB, Preygel IA, Preygel SI. Recombination Variability and Evolution. London: Chapman & Hall; 1994.
66. Noor MA, Grams KL, Bertucci LA, Reiland J. Proc Natl Acad Sci USA. 2001;98:12084–12088. [PMC free article] [PubMed]
67. Whitlock MC, Phillips PC, Moore FBG, Tonsor S. Ann Rev Ecol Syst. 1995;26:601–629.
68. Dahlgaard J, Loeschcke V. Heredity. 1997;78:410–416. [PubMed]
69. Hoffmann AA, Parsons PA. Genetics. 1989;122:837–845. [PMC free article] [PubMed]
70. Rose MR, Vu LN, Park SU, Graves JL. Exp Gerontol. 1992;27:241–250. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...