Evaluation of the rate of evolution in natural populations of guppies (Poecilia reticulata)

Science

Volumn 275, Issue 5308, 1997, Pages 1934-1937

  Reznick, David N ^a   Shaw, Frank H ^b   Rodd, F Helen ^c   Shaw, Ruth G ^b  

^a UNIVERSITY OF CALIFORNIA   (United States)

^b UNIVERSITY OF MINNESOTA   (United States)

^c UNIVERSITY OF CALIFORNIA   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

AGE; ANIMAL EXPERIMENT; ARTICLE; CHILDBIRTH; CONTROLLED STUDY; EVOLUTION; FEMALE; FISH; GENETIC VARIABILITY; MALE; MATURITY; NATURAL SELECTION; NONHUMAN; PHENOTYPE; POPULATION; PREDATOR; PRIORITY JOURNAL; PROGENY; Y CHROMOSOME;

POECILIA RETICULATA;

EVOLUTION; NATURAL SELECTION; TRINIDADIAN GUPPY;

TRINIDAD AND TOBAGO, TRINIDAD;

EID: 0030617889     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.275.5308.1934     Document Type: Article

References (40)

1. C. P. Haskins, E. F. Haskins, J. J. A. Hewitt, in Vertebrate Speciation, F. Blair, Ed. (Univ. of Texas Press, Austin, TX, 1961), pp. 320-395; N. R. Liley and B. H. Seghers, in Function and Evolution in Behavior, G. P. Baerends, C. Beer, A. Manning, Eds. (Oxford Univ. Press, Oxford, 1975), pp. 92-118; J. A. Endler, Evol. Biol. 11, 319 (1978); H. T. Mattingly and M. J. Butler IV, Oikos 69, 54 (1994).
2. C. P. Haskins, E. F. Haskins, J. J. A. Hewitt, in Vertebrate Speciation, F. Blair, Ed. (Univ. of Texas Press, Austin, TX, 1961), pp. 320-395; N. R. Liley and B. H. Seghers, in Function and Evolution in Behavior, G. P. Baerends, C. Beer, A. Manning, Eds. (Oxford Univ. Press, Oxford, 1975), pp. 92-118; J. A. Endler, Evol. Biol. 11, 319 (1978); H. T. Mattingly and M. J. Butler IV, Oikos 69, 54 (1994).
3. C. P. Haskins, E. F. Haskins, J. J. A. Hewitt, in Vertebrate Speciation, F. Blair, Ed. (Univ. of Texas Press, Austin, TX, 1961), pp. 320-395; N. R. Liley and B. H. Seghers, in Function and Evolution in Behavior, G. P. Baerends, C. Beer, A. Manning, Eds. (Oxford Univ. Press, Oxford, 1975), pp. 92-118; J. A. Endler, Evol. Biol. 11, 319 (1978); H. T. Mattingly and M. J. Butler IV, Oikos 69, 54 (1994).
4. C. P. Haskins, E. F. Haskins, J. J. A. Hewitt, in Vertebrate Speciation, F. Blair, Ed. (Univ. of Texas Press, Austin, TX, 1961), pp. 320-395; N. R. Liley and B. H. Seghers, in Function and Evolution in Behavior, G. P. Baerends, C. Beer, A. Manning, Eds. (Oxford Univ. Press, Oxford, 1975), pp. 92-118; J. A. Endler, Evol. Biol. 11, 319 (1978); H. T. Mattingly and M. J. Butler IV, Oikos 69, 54 (1994).
5. J. A. Endler, Environ. Biol. Fishes 9, 173 (1983).
6. D. N. Reznick, M. J. Butler IV, F. H. Rodd, P. N. Ross, Evolution 50, 1651 (1996).
7. D. N. Reznick and J. A. Endler, ibid. 36, 160 (1982); D. N. Reznick, ibid. 43, 1285 (1989); _, F. H. Rodd, M. Cardenas, Am. Nat. 147, 319 (1996).
8. D. N. Reznick and J. A. Endler, ibid. 36, 160 (1982); D. N. Reznick, ibid. 43, 1285 (1989); _, F. H. Rodd, M. Cardenas, Am. Nat. 147, 319 (1996).
9. D. N. Reznick and J. A. Endler, ibid. 36, 160 (1982); D. N. Reznick, ibid. 43, 1285 (1989); _, F. H. Rodd, M. Cardenas, Am. Nat. 147, 319 (1996).
10. D. N. Reznick, Evolution 36, 1236 (1982); _ and H. A. Bryga. Am. Nat. 147, 339 (1996).
11. D. N. Reznick, Evolution 36, 1236 (1982); _ and H. A. Bryga. Am. Nat. 147, 339 (1996).
12. M. Gadgil and P. W. Bossert, Am. Wat. 104, 1 (1970); R. Law, ibid. 114, 399 (1979); R. Michod, ibid. 113, 531 (1979); B. Charlesworth, Evolution in Age Structured Populations (Cambridge Univ. Press, Cambridge, 1994); J. Kozlowsky and J. Uchmansky, Evol. Ecol. 1, 214 (1987).
13. M. Gadgil and P. W. Bossert, Am. Wat. 104, 1 (1970); R. Law, ibid. 114, 399 (1979); R. Michod, ibid. 113, 531 (1979); B. Charlesworth, Evolution in Age Structured Populations (Cambridge Univ. Press, Cambridge, 1994); J. Kozlowsky and J. Uchmansky, Evol. Ecol. 1, 214 (1987).
14. M. Gadgil and P. W. Bossert, Am. Wat. 104, 1 (1970); R. Law, ibid. 114, 399 (1979); R. Michod, ibid. 113, 531 (1979); B. Charlesworth, Evolution in Age Structured Populations (Cambridge Univ. Press, Cambridge, 1994); J. Kozlowsky and J. Uchmansky, Evol. Ecol. 1, 214 (1987).
15. M. Gadgil and P. W. Bossert, Am. Wat. 104, 1 (1970); R. Law, ibid. 114, 399 (1979); R. Michod, ibid. 113, 531 (1979); B. Charlesworth, Evolution in Age Structured Populations (Cambridge Univ. Press, Cambridge, 1994); J. Kozlowsky and J. Uchmansky, Evol. Ecol. 1, 214 (1987).
16. M. Gadgil and P. W. Bossert, Am. Wat. 104, 1 (1970); R. Law, ibid. 114, 399 (1979); R. Michod, ibid. 113, 531 (1979); B. Charlesworth, Evolution in Age Structured Populations (Cambridge Univ. Press, Cambridge, 1994); J. Kozlowsky and J. Uchmansky, Evol. Ecol. 1, 214 (1987).
17. D. N. Reznick and H. Bryga, Evolution 41, 1370 (1987).
18. _ and J. A. Endler, Nature 346, 357 (1990).
19. D. S. Falconer, An Introduction to Quantitative Genetics (Wiley, New York, 1989).
20. J. B. S. Haldane, Evolution 3, 51 (1948); S. C. Stearns, The Evolution of Life Histories (Oxford Univ. Press, Oxford, 1992), pp. 115-118.
21. J. B. S. Haldane, Evolution 3, 51 (1948); S. C. Stearns, The Evolution of Life Histories (Oxford Univ. Press, Oxford, 1992), pp. 115-118.
22. P. D. Gingerich, Science 222, 159 (1983).
23. The 4-year assay was conducted at two levels of food availability. There was an interaction between food availability and population, such that the difference between the control and experimental sites was more pronounced at low levels of food availability (7). The 7.5-year assay was conducted at a single, high level of food availability, which resulted in compressed differences in male size and age, as one would predict from the interaction found in the 4-year assay. An assay run 9 years after the introduction with conditions similar to the 4-year assay produced results that were similar but without the interaction between locality and food availability. Therefore, we conclude that male size at maturity had plateaued 4 years after the introduction was made. The comparisons between the control and introduction populations in the 9-year assay were, respectively, as follows [mean(1 standard error)]: male age, 57.4(1.5) versus 67.6(1.1) days; male size, 62.4(1.5) versus 70.4(1.1) mg; female age, 84.0(1.5) versus 86.2(1.2) days; and female size, 130.9(3.7) versus 129.2(3.0) mg. The control and experimental populations differed significantly for male age and size at maturity. The probability that the female ages differed was marginally significant (P = 0.08), whereas the female sizes did not differ significantly.
24. Guppies breed throughout the year and have overlapping generations (2). We estimated generation time T using the life table method [R. E. Ricklefs, Ecology (Chiron, Newton, MA, 1973)], which is based on estimates of birth and death schedules in natural populations. We estimated mortality rates from mark-recapture experiments on guppies from seven low-predation localities, including the Aripo and El Cedro experimental sites (3). Because the life table method requires age-specific survival, we estimated the ages of individuals in different size classes using growth rate equations derived for the same fish in the mark-recapture experiments (3) based on their sizes at the beginning and end of the recapture interval. We estimated size-specific fecundity with dissections and embryo counts from females collected from these same localities, plus from additional collections (4). We estimated the frequency of reproduction from laboratory studies, conducted at temperatures and food availabilities similar to those in the field (5). With this information, we estimated the mean generation time of fish in low-predation localities to be 210 days, or 1.74 generations per year.
25. R. Lande, Evolution 33, 402 (1979).
26. T. Mousseau and D. Roff, Heredity 59, 181 (1987).
27. K. D. Kallman, in Ecology and Evolution of Livebearing Fishes (Poeciliidae), G. K. Meffe and F. F. Snelson, Eds. (Prentice Hall, Rivers Edge, NJ, 1989), pp. 163-184; J. Travis, in Ecological Genetics, L. A. Real, Ed. (Princeton Univ. Press, Princeton, NJ, 1994), pp. 205-232.
28. K. D. Kallman, in Ecology and Evolution of Livebearing Fishes (Poeciliidae), G. K. Meffe and F. F. Snelson, Eds. (Prentice Hall, Rivers Edge, NJ, 1989), pp. 163-184; J. Travis, in Ecological Genetics, L. A. Real, Ed. (Princeton Univ. Press, Princeton, NJ, 1994), pp. 205-232.
29. R. Lande and S. J. Arnold, Evolution 37, 1210 (1983).
30. B. Charlesworth, R. Lande, M. Slatkin, ibid. 36, 474 (1982).
31. R. Lenski and M. Travisano, Proc. Natl. Acad. Sci. U.S.A. 91, 6808 (1994).
32. H. L. Gibbs and P. R. Grant, Nature 327, 511 (1987).
33. S. J. Gould and N. Eldredge, ibid. 366, 223 (1993).
34. SAS/STAT User's Guide: Release 6.03 Edition (SAS Institute, Cary, NC, 1988), pp. 433-506.
35. H. D. Patterson and R. Thompson, Biometrika 58, 545 (1971); K. Meyer, J. Dairy Sci. 66, 1988 (1983); R. G. Shaw, Evolution 41, 812 (1987).
36. H. D. Patterson and R. Thompson, Biometrika 58, 545 (1971); K. Meyer, J. Dairy Sci. 66, 1988 (1983); R. G. Shaw, Evolution 41, 812 (1987).
37. H. D. Patterson and R. Thompson, Biometrika 58, 545 (1971); K. Meyer, J. Dairy Sci. 66, 1988 (1983); R. G. Shaw, Evolution 41, 812 (1987).
38. Guppies from the experimental site on the El Cedro River (7.5-year assay) and the control site on the Aripo River (11-year assay) were reared in a three-generation, extended-pedigree, paternal half-sibling design. Paternal half-sibling matings were performed for the first and second generations of labreared descendants from each locality. The dependent variables (age and size at maturity in males and females) were quantified for the second and third laboratory-born generations. Sample sizes for the Aripo River were 16 wild-caught sires and 30 second-generation, laboratory-reared dams for the first generation cross, yielding data on 49 daughters and 57 sons. Of these daughters, 39 were used as dams for the next generation. They were mated to 23 wild-caught sires that were unrelated to individuals from the first two generations. These crosses yielded data for 115 daughters and 114 sons in the third generation. Sample sizes for the El Cedro River were 15 wild-caught sires and 40 labreared dams for the first generation, yielding data for 78 daughters and 97 sons. Twenty-eight daughters from this generation were mated to 11 unrelated, wild-caught males, yielding data for 73 daughters and 85 sons in the third generation. Sexes were analyzed separately because the genetic variance-covariance matrices indicated strong sex linkage for genetic variation in age and size at maturity.
39. The P and G matrices were sampled from a multivariate normal distribution using the estimates of P and G as means and their sampling variance-covanance matrix as variance. The values for R, the vector of responses to selection, were sampled from a multivariate normal distribution with means equal to H (Table 1) and a sampling variance-covariance matrix obtained from the multivariate analysis of variance that compared the control and experimental populations. Vectors of 1000 values of β and S were calculated, and the distributions of β and S were inferred from these. Some of the sampled G matrices were not positive definite and hence could not be inverted. In these cases, the implied estimates of β and S were either positive or negative infinity, depending on the associated value of R, and were retained as such in our set of 1000 β and S vectors. If more than 25 of the 1000 estimates were positive infinity, the confidence interval was deemed to have no upper bound. Consequently, there are four different ways of reporting the results, depending on the nature of the G matrix and the results of the 1000 simulations: (i) One-sided confidence intervals are reported when singularity of more than 2.5% of the sampled G set the upper confidence limit at infinity. The one-sided value equals the 25th value of 1000 simulations (rank ordered from smallest to largest) and hence is equivalent to the lower bound of a 95% confidence interval, (ii) Confidence intervals with lower and upper bounds are reported when the simulations allowed us to set both an upper and lower limit to the distributions. The reported values are the 25th and 975th values of the 1000 simulations, (iii) Bivariate estimates of β and S were only possible for the males from both experiments. The only data set for which we could estimate the confidence limits for the bivariate analysis was the El Cedro males, for both the 4-and 7.5-year results. Here the limits mark the two-dimensional range of the 950 simulated values closest (in Euclidean distance) to the estimates, (iv) Standard errors were not estimable for the bivariate estimates of β and S for males from the Aripo River because of the near singularity of the G matrix. In this case the difference between β values for age and size at maturity are more pronounced in the bivariate analysis which takes the high genetic correlation between them into account, than in the univariate analysis.
40. We thank B. Brodie and K. Hughes for help with the computation of the selection gradients. All of the laboratory studies were performed by H. Bryga. D.N.R. was supported by NSF grants BSR8818071, DEB-9119432, and DEB-9419823. D.N.R. dedicates this paper to the memory of his father, Mortimer M. Reznick.