Role of mutation in fly-wing evolution

Analysis of wing variation within and between fly species reveals an unexpectedly slow evolutionary rate. Variations due to mutation and interspecific differences are similar, perhaps as a result of complex genetic interactions. See Letter p.447

The origin of macroevolution — the variation between species — has been much debated. Does it arise as a consequence of microevolutionary processes that occur at the species level1, such as mutation, selection and the random genomic changes in a population known as genetic drift? Or does it require additional processes that act above the level of individual species, such as species selection that arises from differences in the rate of formation of new species (because certain species are more likely than others to split and give rise to two species), or in the rate of species extinction2? If such differences in the rate of speciation and extinction can be linked to some species trait, this can give rise to species selection. On page 447, Houle et al.3 provide a remarkable experimental test of whether mutation can account for interspecific differences in wing shape in species of the fruit-fly family Drosophilidae.

The authors gathered an unprecedentedly large array of two-dimensional data on the relative positions at which fly-wing veins intersect at points called landmark coordinates (Fig. 1). In this study and in previous work4,5 by the same group, such data were obtained using an ingenious device that allows the left wings of anaesthetized flies to be immobilized and spread, then digitally photographed6. The wing veins are mathematically modelled as curves, and the landmark coordinates extracted from these models6. Just one minute is needed to collect and process data from a single specimen. This system provides cutting-edge data collection for shape measurements that are both extremely fast and accurate. Such an approach is crucial in obtaining the sample sizes required for the type of study conducted by Houle and colleagues.

Figure 1: Fly-wing evolution.

David Houle

The wing veins of the fruit fly Drosophila melanogaster intersect (red) at positions termed landmark coordinates. Houle et al.3 analysed the intersection positions of wing veins in around 50,000 drosophilid fly wings representing different species and large samples of the same species. This enabled the authors to assess how mutation and natural genetic variation affect variations in the positioning of these landmark coordinates and influence the evolution of species differences. They found that changes in the wing-vein patterns evolve very slowly, given the amount of variation that is produced by mutations. This suggests that a strong stabilizing selection favouring intermediate trait values occurs in all of the species analysed.

The authors assessed the patterns of variation — which traits (in this case, landmark-coordinate positions) vary widely, which vary only a little, which vary together and which vary independently of each other. They used these data to address the relationship between wing-vein variations caused by newly arising mutations, those due to differences between species and those caused by standing genetic variation, which occurs within a species owing to the existence of different versions (alleles) of a gene.

The variation produced by mutation was previously estimated by Houle and colleagues4 using around 12,000 specimens. The authors had measured this variation in mutation-accumulation experiments, generating large numbers of sublines from inbred fly lines, breeding these sublines for many generations and then measuring mutational variation from the wing-vein differences between the sublines4. The researchers also previously estimated the standing genetic variation by analysing the wing-vein similarity between relatives in nearly 17,000 wings from a single natural population of Drosophila melanogaster from Florida5. The wing-vein variation between different species was measured by Houle et al. using about 21,000 specimens from 117 fly species.

Precise quantitative estimates of the variation caused by mutation and standing genetic variation are notoriously difficult to obtain7. The sample sizes required must be orders of magnitude higher than those needed to analyse 'ordinary' variation in form (variation in form that has not been separated into that caused by genetic or environmental variation). The data set now gathered by the authors is exceptional, because it allows measurement of the variation caused by mutation, by standing genetic variation and by the differences between species caused by evolutionary diversification — all in the same system.

The three of kinds of variation estimated by Houle and colleagues are related by quantitative genetic evolutionary theory8,9,10. According to this theory, if natural selection is responsible for variation between species, then this variation is a function of within-species standing genetic variation and variation in directional selection of traits (in which an extreme form is favoured over the others) for fly species in the same clade8. Standing genetic variation is itself a function of new mutational variation and of the pattern and strength of a form of natural selection, known as stabilizing selection9, that favours intermediate values of a trait over those at the extremes.

However, if interspecific variation is caused by genetic drift (random genetic changes), then it would be expected to be proportional to standing genetic variation10, which, in turn, is a function of the rate and pattern of the variation produced by mutation11. In this genetic-drift model, variation produced by mutation results in standing genetic variation, which, in turn, results in interspecific variation. In quantitative genetic evolutionary theory, macroevolutionary patterns of interspecific variation arise from underlying microevolutionary processes. These include: directional selection favouring a particular trend of change; stabilizing selection favouring intermediate trait values over the extremes; new variation produced by mutation; and the levels of standing genetic variation. As is usual in quantitative evolutionary theory, these relationships assume an additive model of gene effects, in which an allele's effects are not affected by other alleles of the same gene, or by alleles of other genes.

One of the authors' key conclusions is that the pattern of interspecific wing-trait variation is remarkably similar to the patterns of wing-trait variation that result from standing genetic variation or from mutation. Such a similarity would be expected if evolution is driven by genetic drift, or if patterns of stabilizing and directional selection are themselves either unstructured or aligned with mutational and genetic-variation patterns.

However, the results that Houle et al. obtain for the levels of variation within and between species, and for the rates of wing-vein evolution, put a fly in the ointment. Changes in wing-vein shape evolve very slowly compared with the level of standing genetic variation or the amount of variation produced by mutations within species over evolutionary time. Over time, the expected evolutionary rate given genetic drift alone is 10,000 times higher than the rate observed by Houle and colleagues. They conclude that evolutionary rates are not constrained by mutation or lack of standing genetic variation.

This suggests that the same strong stabilizing selection is imposed on wing-vein shape in all the species. Most mutations that affect more than one gene and that are not aligned with the direction of evolution would need to be deleterious for the observations on the amount of variation produced by mutation over time to fit the extremely low rate of evolution. Previous work11 came to similar conclusions regarding the slow rate of mammalian brain evolution, given the rate of production of new variation by mutation and the diversification of the species by genetic drift. It seems that evolutionary rates are severely limited by strong stabilizing selection occurring in the same pattern within species in the same clade.

Houle and colleagues raise another intriguing possibility, which is not usually considered. Instead of the model described above, in which new mutational variation in fly-wing shape and strong stabilizing selection influence interspecific variation, the variation due to mutation itself evolves. In this situation, the similarity between mutational and interspecific-variation patterns occurs because both are responding to the same directional-selection pressures.

The theoretical basis for this alternative scenario is still in its infancy, and much additional research will be needed to flesh it out12,13,14,15,16. It has been shown13 that, in the presence of non-additive interactions between genes, the effect of a specific gene is itself variable and can evolve. Under directional selection, a gene's effect on the trait being selected for can increase because of changes in allele frequencies at other genes. This increase in effect is reflected in the overall variation of the trait. The extent and pattern of mutational variation evolve to match the adaptive requirements of directional selection. In this scenario, the similarity between mutational and interspecific-variation patterns lies in their common response to directional-selection patterns.

Houle and colleagues provide an example in which the macroevolutionary pattern of interspecific differences in fly wings is due to a balance between microevolutionary processes. Whether such a pattern occurs in other species will require further study. However, it will probably be difficult to find an animal that can be used as effectively as this one to marshal the evidence for mutational variation, standing genetic variation and interspecific variation in the same system.Footnote 1


  1. 1.

    See all news & views


  1. 1

    Futuyma, D. J. in Macroevolution: Explanation, Interpretation and Evidence (eds Serrelli, E. & Gontier, N.) 29–85 (Springer, 2015).

    Google Scholar 

  2. 2

    Gould, S. J. Science 216, 380–387 (1982).

    CAS  Article  ADS  Google Scholar 

  3. 3

    Houle, D., Bolstad, G. H., van der Linde, K. & Hansen, T. F. Nature 548, 447–450 (2017).

    CAS  Article  ADS  Google Scholar 

  4. 4

    Houle, D. & Fierst, J. Evolution 67, 1116–1130 (2013).

    Article  Google Scholar 

  5. 5

    Mezey, J. G. & Houle, D. Evolution 59, 1027–1038 (2005).

    Article  Google Scholar 

  6. 6

    Houle, D., Mezey, J., Galpern, P. & Carter, A. BMC Evol. Biol. 3, 25 (2003).

    Article  Google Scholar 

  7. 7

    Cheverud, J. M. Evolution 42, 958–968 (1988).

    Article  Google Scholar 

  8. 8

    Felsenstein, J. Annu. Rev. Ecol. Syst. 19, 445–471 (1988).

    Article  Google Scholar 

  9. 9

    Lande, R. Genetics 94, 203–215 (1980).

    CAS  PubMed  PubMed Central  Google Scholar 

  10. 10

    Lande, R. Evolution 33, 402–416 (1979).

    Article  Google Scholar 

  11. 11

    Lynch, M. Am. Nat. 136, 727–741 (1990).

    Article  ADS  Google Scholar 

  12. 12

    Cheverud, J. M. J. Theor. Biol. 110, 155–171 (1984).

    CAS  Article  Google Scholar 

  13. 13

    Pavličev, M. & Cheverud, J. M. Annu. Rev. Ecol. Evol. Syst. 46, 413–434 (2015).

    Article  Google Scholar 

  14. 14

    Draghi, J. & Wagner, G. P. Evolution 62, 301–315 (2008).

    Article  Google Scholar 

  15. 15

    Jones, A. G., Bürger, R. & Arnold, S. J. Nature Commun. 5, 3709 (2014).

    CAS  Article  ADS  Google Scholar 

  16. 16

    Melo, D. & Marroig, G. Proc. Natl Acad. Sci. USA 112, 470–475 (2015).

    CAS  Article  ADS  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to James Cheverud.

Related links

Related links

Related links in Nature Research

Evolution: Differences can hold populations together

Evolution: Catastrophe triggers diversification

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Cheverud, J. Role of mutation in fly-wing evolution. Nature 548, 401–403 (2017).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.